Michlet 9.20 Released

[Remmlinger]

Originally Posted by Leo Lazauskas

There is some work on surface-piercing foils at the IIHR:
http://old.iihr.uiowa.edu/~shiphydro/efd.htm

I am still very wary of CFD results for Wigley hulls. Some cherry-pick experiments that give the best "agreement", others tweak the parameters in their codes to get good agreement with experiments. Others are just not very good, e.g. as shown in:
Michell v CFD and EFD: Round 2007
That Wigley hull has L/B = 0.8 and it can be seen that the CFD code is pretty poor, as is thin-ship theory. It also took the CFD code about 7 days on several processors to calculate the drag at one speed!

I hope Openfoam is better, but I'd like to know how many parameters they tweaked to get the agreement. When I know that, I'll know how many I can adjust in some of my codes, e.g.
Michell-Lamb Wave Resistance Integral for Thin Ships

There I choose the value of one parameter, a non-dimensional "viscosity".

Leo.

Thank you for the helpful links.
It seems that mentioning CFD is like touching your red button, but I guess I get your point.
I am not very keen to spend many months to get a CFD software running, only to find out that my computer is too slow and the results are unreliable. I wish there were an alternative.
For mathematicians the world is easy. They develop a calculation method and then they look for an application that suits this method. No doubt the applause for Michlet is well deserved, it is the ideal tool for thin ships.
In contrast engineers are first confronted with a problem and this starts the search for a suitable tool. This is my situation. All VPPs that I know are lacking a realistic estimate of the additional wave resistance caused by the keel. Adding a lift producing keel will surely change the wave resistance of the yacht compared to the bare hull (you mentioned the AMECRC paper by Binns). To develop an estimate for this additional drag one needs the drag force of a yacht with and without keel at heel and leeway. The yacht in this case is not a thin ship. Which tool would you use? I tried to solve it experimentally in a circulating water channel, but the experimental error was larger than the drag differences I was looking for.
Uli

[Leo Lazauskas]

Originally Posted by Remmlinger

Thank you for the helpful links.
It seems that mentioning CFD is like touching your red button, but I guess I get your point.
I am not very keen to spend many months to get a CFD software running, only to find out that my computer is too slow and the results are unreliable. I wish there were an alternative.
For mathematicians the world is easy. They develop a calculation method and then they look for an application that suits this method. No doubt the applause for Michlet is well deserved, it is the ideal tool for thin ships.
In contrast engineers are first confronted with a problem and this starts the search for a suitable tool. This is my situation. All VPPs that I know are lacking a realistic estimate of the additional wave resistance caused by the keel. Adding a lift producing keel will surely change the wave resistance of the yacht compared to the bare hull (you mentioned the AMECRC paper by Binns). To develop an estimate for this additional drag one needs the drag force of a yacht with and without keel at heel and leeway. The yacht in this case is not a thin ship. Which tool would you use? I tried to solve it experimentally in a circulating water channel, but the experimental error was larger than the drag differences I was looking for.
Uli

I have no problem with CFD if it is the appropriate tool for the job. And I do like the work of Prof. Fred Stern and his group because they are very much committed to proper validation and verification.

Openfoam requires a lot of programming, so it is not an easy CFD code for beginners. I do like the fact that it is open source though.

I'm not sure what you could try instead of CFD. Potential codes using Neumann-Kelvin type boundary conditions should be better than thin-ship theory, but then there are still problems with consistency. Satisfying one boundary condition better, but still linearising another is not consistent and will improve estimates in many cases, but not in many others. Fully non-linear codes (like RAPID, Shipflow etc) are not guaranteed to find a solution in all cases, especially for vessels with large transoms. Some just use linear theory when they fail, which is not much better than using thin-ship theory in many cases.

My late colleague Ernie Tuck used to say that ship hydrodynamics was about 50 years behind aerodynamics. I don't think that situation has changed much in the last 15 years.

Good luck!
Leo.

[thudpucker]

Me (the software inept) downloaded Micolet from the link in the first post.
I thought since it was free, I'd try my hand at drawing a boat. I still cannot get it to work.

Suddenly I feel like I'm trying to do Neurosurgery from a DIY sketch.
After I downloaded it, what was I supposed to do next?
Who knows what I'm supposed to do now?

[FMS]

You should have a /docs directory containing michman_920.pdf which is a concise manual

Michlet is for analysis not drawing. You export - input your hull shape definition from freeship, delftship, rhino, etc. to analyze
Freeship to Michlet questions

[thudpucker]

Thanks FMS, I did not realize it was an eval program. I'll delete it. I was looking for an easy boat drawing program.

[patrik111]

Hello,

Is there a way to run Michlet and Godzilla under 64 bit windows 7?
(or any other, linux?, 64 bit OS)

I have tried compatibility mode and running as administrator, but with little success. (the program runs, but quite is not very responsive to keystrokes)



Kind regards

Patrik Elfving

[Leo Lazauskas]

Originally Posted by patrik111

Hello,

Is there a way to run Michlet and Godzilla under 64 bit windows 7?
(or any other, linux?, 64 bit OS)

I have tried compatibility mode and running as administrator, but with little success. (the program runs, but quite is not very responsive to keystrokes)

Kind regards

Patrik Elfving

Michlet/GODZILLA works fine under Windows 7 on several machines I have tried. I didn't need to use "compatibility mode".

It is not available for other OS at present.

Good luck!
Leo.

[The Scuttlebut]

Hi all,

I've been using Michlet and trying to compare it to other methods we use for resistance estimation, to see if we can add it to our stable of methods. I am finding that at froude numbers from 0.05 to around 0.3, there is a sharp increase in resistance predicted from Michlet, then a plateau. This doesn't appear in the other methods (Holtrop, Feng-Leibman, etc.). I compared it to my thesis model test data for a long slender hull (268m long 25m beam, 9m draft) and found the same issue with real data (full scale using the ITTC 57 method). Anyone have any insight as to this? On displacement and semi-planing/semi-displacement hulls I see the same character to the resistance curve.

Attached find a graph of the comparison of my thesis data to various methods.

[DCockey]

Originally Posted by The Scuttlebut

Hi all,

I've been using Michlet and trying to compare it to other methods we use for resistance estimation, to see if we can add it to our stable of methods. I am finding that at froude numbers from 0.05 to around 0.3, there is a sharp increase in resistance predicted from Michlet, then a plateau. This doesn't appear in the other methods (Holtrop, Feng-Leibman, etc.). I compared it to my thesis model test data for a long slender hull (268m long 25m beam, 9m draft) and found the same issue with real data (full scale using the ITTC 57 method). Anyone have any insight as to this? On displacement and semi-planing/semi-displacement hulls I see the same character to the resistance curve.

Attached find a graph of the comparison of my thesis data to various methods.

Have you run the examples which are included in the download? Do the results from them look reasonable?

[Leo Lazauskas]

Originally Posted by The Scuttlebut

Hi all,

I've been using Michlet and trying to compare it to other methods we use for resistance estimation, to see if we can add it to our stable of methods. I am finding that at froude numbers from 0.05 to around 0.3, there is a sharp increase in resistance predicted from Michlet, then a plateau. This doesn't appear in the other methods (Holtrop, Feng-Leibman, etc.). I compared it to my thesis model test data for a long slender hull (268m long 25m beam, 9m draft) and found the same issue with real data (full scale using the ITTC 57 method). Anyone have any insight as to this? On displacement and semi-planing/semi-displacement hulls I see the same character to the resistance curve.

Attached find a graph of the comparison of my thesis data to various methods.

Sorry, but I can't help much without seeing the input file.
Does this hull have a large transom stern?

I haven't seen that sort of behaviour except for small models of hulls with large transom sterns.
Or when people have not produced a correct input file of offsets (e.g. putting the ship back to front, or upside down).

Is the pointy end at the front in the offset section?

Good luck!
Leo.

[The Scuttlebut]

The lines plan in Michlet looks right.

The boats in question do have large transom sterns. I would love to provide the input file, but although it was my thesis, I am unsure of my ability to share the hullform due to the people who created it. I am assuming from your question that the transom stern could be the reason for the hump in resistance?

I ran another design we have with a small transom and got better results, I assumed at the time it was because of the displacement hull type, as opposed to the semi-displacement hulls I am currently working on.

Any ideas how to counter act this or is it just a limitation of the theory used in the program?

Thanks.

[Leo Lazauskas]

Originally Posted by The Scuttlebut

The lines plan in Michlet looks right.

The boats in question do have large transom sterns. I would love to provide the input file, but although it was my thesis, I am unsure of my ability to share the hullform due to the people who created it. I am assuming from your question that the transom stern could be the reason for the hump in resistance?

I ran another design we have with a small transom and got better results, I assumed at the time it was because of the displacement hull type, as opposed to the semi-displacement hulls I am currently working on.

Any ideas how to counter act this or is it just a limitation of the theory used in the program?

Thanks.

It is a limitation of the thin-ship theory implemented in Michlet.
Flotilla might give better results, but I'd need the offsets to run that.

Experiments with large transoms at low Froude numbers can also be very prone to uncertainty.
This is something that the ITTC is trying to address because Prohaska's method for determining
form factors requires accurate experimental values of resistance at low Fr,
but this is exactly where the greatest experimental uncertainties occur.

On the other hand, judging by your graph, Michlet gives reasonable results for Fr > 0.3.
Should anyone be seriously concerned with the performance of ships with large transoms at low Fr?

Good luck!
Leo.

[patrik111]

Hi Leo,

Donation made!

I have now tried this on several machines and I am getting the same symptoms. i.e. the programs runs but extremely slowly.

Today I have a vista sp2 32 bit computer, but this is going to be a cnc router controller in a, hopefully, not to distant future. If there is one piece of software I would like to transfer to more up to date computers, it is michlet/godzilla. I have been a fan since I discovered it around v8,07. But now I am having no success as of late.

Would you have any idea on what might be causing this?
Anyone else have had simalar issues ?
is possible to re-compile the program under win7 64?

Kind regards

Patrik Elfving


Kind regards
Patrik Elfving

[Leo Lazauskas]

Originally Posted by patrik111

Hi Leo,

Donation made!

I have now tried this on several machines and I am getting the same symptoms. i.e. the programs runs but extremely slowly.

Today I have a vista sp2 32 bit computer, but this is going to be a cnc router controller in a, hopefully, not to distant future. If there is one piece of software I would like to transfer to more up to date computers, it is michlet/godzilla. I have been a fan since I discovered it around v8,07. But now I am having no success as of late.

Would you have any idea on what might be causing this?
Anyone else have had simalar issues ?
is possible to re-compile the program under win7 64?

Thanks Patrick!

I tried Michlet on two different i7 computers using the Godzilla example input file gpenta58.mlt

This should be a very severe test for any machine because it tries to "optimise" the dimensions, shapes and placement of 5 independent hulls.

It took 30 seconds to evaluate 4706 combinations and I got identical results under Windows 7 and XP.

Could someone else please try the same input file for comparison?

Good luck,
Leo.

[aki123]

First: Michlet is great! Thanks for sharing!

For running Michlet in Linux (Ubuntu) do the following:

.) Install 'wine' - the windows emulator
.) Extract the mlt920.zip to some directory
.) Run Michlet via doubleclick or terminal command ./mlt920.exe

So Michlet is running in wine. It is running fine unless you try to move the window. Then wine crashes. So keep the window positioned, where it pops up.

Scripting Michlet:
I tried to feed Michlet via a command-file that contains all the keyboard operations I need to get the desired analysis output: ./mlt920.exe < input.txt
With input.txt containing the letters for controlling Michlet. Unfortunately this seems not to work. I will continue on this.

All the best
Aki