Michlet for retroactive wave-generation

[Metatron]

Hy;

I got some general question about the validity of Michelet-results in respect to a real world build ship. Basically I retroactively calculate the wave-pattern based on the ship, in this case it's the "Bayern".

I read a lot of the threads here that contain something about Michlet, and I'm clear that the software is usefull only under contrains, so I'd like to ask if you think the given ship here is within those limits, and if you think the generated wave-pattern is genuine.

I added three photos of nice shots where the wave-pattern is really clearly visible. I do not know the speeds of the ship on the particular photos, but I think it should be regular marsh which is about 20kn.

I detected that Mitchel does not calculate the bow-wave, or let's say there is no data (all 0s) to the left of 0.0. Is that normal? In the manuals of Michel the submarine-shot at least have a very nice bow-wave. Also the ship-hull itself is positioned as red object in the shots. I suppose this is done manually? Or is there a shortcut to add the hull? Sadly I had to rely on the spanish manual, the english was nowhere to be found (I speak spanish, but the topic is very technical).

And well, my final question would be how I can handle the resulting .mlt file with the wave-pattern grid. I may programm a converter to .obj, but if there is a solution available I'd take that first.

Ah, final final question. How big is the impact of the natural ocean waves on the generated wave-pattern. I suppose there is some complex damping going on. Or do you think it's okayish to simple accumulate ocean-wave to the wave-pattern. Do I have to +- accumulate in respect to the ocean plane? (I mean if below-plane ocean-waves and below-plane wake-waves amplify to even lower values?)

Thanks
Niels

[apex1]

Very interesting questions! Thank you.
And I guess you started a thread worth looking at!

Because your questions will not be answered in a straight pattern (that is not possible), but in straight announcements! sic....



is there a Willoughby around?.........

[Ad Hoc]

Before delving in too deep, what is the purpose of wanting to calculate the wave pattern, or rather, have a program calculate the wave pattern (rather than by hand calculations)?

The objective is..?

[Metatron]

Creating the thing in virtuality of course. Ship's gone. ;^)

I created the hull in FreeShip as good as I could in respect to the ship-lines I have, the measures are all correct (volumen, length, beam etc.). I don't know if Mitchlet could handle the rudders, or the propeller.

I ask the question, because I could make 10 other hulls of the same generation (10s to 20s) which are quite distinct, like less beam, less depth etc. If the Michlet-results would be bogus I would probably not create the hulls with FreeShip (which is SDS) but rather with Lathe-NURBS, and have to accept that I have to make the wave-pattern by hand, I mean drawing and deforming them by hand so they look nice ... or I just copy any Kevin-waveform and stretch them under the ship's body, which is kinda lame ...

It would make me happy if I could recreate those ships not only by appearance, but also re-creating their physical behaviour and impact.

Ciao
Niels

[Ad Hoc]

Niels
There are really only two ways in which can achieve your objective:
1) Visually, from the photo's you ahve posted
or
2) By taking, or acquiring, full scale trial measurements and then calibrating the results accordingly.

Not an easy task, whichever route you take...

[apex1]

It would make me happy if I could recreate those ships not only by appearance, but also re-creating their physical behaviour and impact.
Ciao
Niels

Yeah, that would be a nice game, but worth?
First you should know that you DID NOT create them in Freeship!
It is a mistake to believe one could make a ancient hull design happen by some input of data!
The next mistake is: to believe one would make a "wave pattern" (related to real worlds data) out of that!
If one studies the life of Naval Architecture, one does notice pretty soon, it is not pushing buttons! If one goes out to sea, one does notice pretty soon why.............................

Regards
Richard

[Metatron]

Yeah, that would be a nice game, but worth?
First you should know that you DID NOT create them in Freeship!
It is a mistake to believe one could make a ancient hull design happen by some input of data!
The next mistake is: to believe one would make a "wave pattern" (related to real worlds data) out of that!
If one studies the life of Naval Architecture, one does notice pretty soon, it is not pushing buttons! If one goes out to sea, one does notice pretty soon why.............................

Regards
Richard

Uhm, worth? If I need 7 days deforming a mesh, or if I need 1 hour to get a .mlt, makes a difference. Clicking together a sensefull SDS, or virtually soldering steel-plates together within a physic based dynamics system, makes also a difference. I don't intent to recreate *the* reality^tm, I just want to get an idea about how (visually or functional) acceptable the retroactive approach to wave-generation is. In principle and in particular. And when and where it fails, or looks obviously wrong, how to work it out better. I know very little about hydrodynamics BTW.

For example in the following shots, the first ist "Indomitable" and has a very characteristic wave-pattern. Can this be reproduced?

And the second is "König", the predecessor class of the "Bayern" with pretty much just proportional changes, why is the wave-pattern so narrow? Michlet give very broad patterns (probably because of the slow speed of 20kn), but those 22kt monsters have relative narrow wave-patterns. Which parameter is wrong in the calculation? Is it the water-inertia which is not modeled in Michlet? Can I fake inertia by just making the ship steam twice as fast as it should?

If it looks great and you know it's also "sort of" realistic, doesn't make it the necessary adjustments worth it? : ^)

Thanks
Niels

[Ad Hoc]

Niels

Ah ok, i see where you are coming from now. You have no training/background/education in hydrodynamics. That complicates things a little since there is much the learn.
Perhaps you should read post #3 in the link here:
http://www.boatdesign.net/forums/boat-design/wave-pattern-question-28667.html
this shall provide you with a very basic grounding on wave patterns.

[Metatron]

Niels
There are really only two ways in which can achieve your objective:
1) Visually, from the photo's you ahve posted
or
2) By taking, or acquiring, full scale trial measurements and then calibrating the results accordingly.

Not an easy task, whichever route you take...

Right. As I don't know so much about HD, I for example could not exactly say which difference (of photo to simulation) is caused by what. I guessed you (all) have maybe a better intuition.

For example what makes the wave-pattern narrower, if not the speed? Is a suggestion of an adjustment possible?

That's why I posted the shots. I suppose you may even be much more sensible to differences/inconsistancies than me ...
Personally I'm quite surprised how reasonable the Michlet results are, visually.

Ciao
Niels

[Rick Willoughby]

......
Personally I'm quite surprised how reasonable the Michlet results are, visually.

Ciao
Niels

Niels
Why would it surprise you? It accurately models the wave energy.

The author of Michlet, Leo Lazauskas, has been contemplating working backwards from the wave pattern to create the hull shape.

From my observations of smaller hulls in the range to about 10m the wave pattern produced by Michlet is accurate.

Michlet will be more accurate for open water wave prediction than data from tank testing because the tank boundaries are too close to accurately reproduce the far-field boundary conditions of a deep ocean.

Rick

[Leo Lazauskas]

Niels
The author of Michlet, Leo Lazauskas, has been contemplating working backwards from the wave pattern to create the hull shape.


That capability has been in Michlet for many versions.

Remember too, Michlet is reasonably accurate for thin hulls in calm water. Ambient waves and wind, irregular bathymetry, variable eddy viscosity, non-constant speed, and many other factors will all contribute to distorting the ideal far-field wave patterns predicted by Michlet.

Have fun!
Leo.

[Metatron]

That capability has been in Michlet for many versions.

Remember too, Michlet is reasonably accurate for thin hulls in calm water. Ambient waves and wind, irregular bathymetry, variable eddy viscosity, non-constant speed, and many other factors will all contribute to distorting the ideal far-field wave patterns predicted by Michlet.

Have fun!
Leo.

Hy Leo; thanks for being interested in this thread;

Yes, I understood so much that the (simulated) environment is artefact-free. What would be your recommendation to mix the wave-pattern with the ocean-plane patterns?

Maybe you could be so kind to give me a hint to these software issue:
- no wave-pattern left of 0.0 (but your submarine shots show data left of 0.0)
- seems the grid limit is at something near 300x300, you can confirm that?

Well and I guess it's hard to tell, but you got an idea about the reason for that some of the "real" wave-patterns on the photos are more narrow? Is this environment caused, or am I starting to use Michlet beyond it's specifications? I mean those are no little hulls ... and they have high weight/length ratio (having 250mm steel-belt and all, ca. 11kt).

You maybe also got another bit of knowledge about the significance of the rudder and propeller forces? Do they contribute? With those ships there is often a much ship-line parallel foam-line leaving the back. Is this wave-pattern related or is this completely based on turbulence generated at the back either by the ship's body or the propellers?

Thank you much
Niels

[Leo Lazauskas]

Yes, I understood so much that the (simulated) environment is artefact-free. What would be your recommendation to mix the wave-pattern with the ocean-plane patterns?


I guess you could try to add some form of random ambient waves with amplitudes and directions taken from some standard spectrum. Or you could use long-crested sine waves with or without a random component. It depends on what you are trying to mimic.


Maybe you could be so kind to give me a hint to these software issue:
- no wave-pattern left of 0.0 (but your submarine shots show data left of 0.0)
- seems the grid limit is at something near 300x300, you can confirm that?


As the manual says... Michlet only produces far-field waves, i.e. aft of the stern. The submarine patterns were produced with other code that includes near-field effects.


Well and I guess it's hard to tell, but you got an idea about the reason for that some of the "real" wave-patterns on the photos are more narrow? Is this environment caused, or am I starting to use Michlet beyond it's specifications? I mean those are no little hulls ... and they have high weight/length ratio (having 250mm steel-belt and all, ca. 11kt).


The size of the ship doesn't matter to Michlet. (In pure inviscid theory it is impossible to determine both the length of the ship and its speed from the wave pattern.)

The effective viscosity (e.g. molecular + eddy) behind the hull tends to damp out short wavelength waves. If the ship is travelling at a Froude number where short (i.e. diverging waves) predominate, then that viscosity might make the pattern appear narrower by damping waves travelling at certain wave angles. Sorry, but there are just too many unknowns here to make a plausible guess.



You maybe also got another bit of knowledge about the significance of the rudder and propeller forces? Do they contribute? With those ships there is often a much ship-line parallel foam-line leaving the back. Is this wave-pattern related or is this completely based on turbulence generated at the back either by the ship's body or the propellers?


It is probably a combination of the turbulence created by the props and appendages, and also by the boundary layer as it separates from the hull near the stern. These, of course, interact with the waves produced by the ship and the ambient waves. If the ship is pitching, heaving and rolling, that can also significantly affect the wave wake.

You are definitely pushing Michlet too much if you are trying to simulate the wave pattern close to the ship. That is not what the code is for.

Good luck!
Leo.

[Leo Lazauskas]

Niels
Why would it surprise you? It accurately models the wave energy.
Rick

A very powerful check on the (inviscid) code is to generate wave elevations along a (very long!) wave cut parallel to the ship's track, and then use Fourier transforms to get the so-called P, Q functions. These can then be integrated to get the wave resistance. That's a fairly standard technique used by towing tanks, and it is also a reasonable method for some situations such as yawed pressure distributions where other numerical techniques are computationally too messy.

Yet another check (not possible in Michlet because it doesn't include the near field) is to use the pressures on the hull to calculate the wave resistance.

Cheers,
Leo.

[Metatron]

As the manual says... Michlet only produces far-field waves, i.e. aft of the stern. The submarine patterns were produced with other code that includes near-field effects.

So I guess right that 0.0 is the middle of the ship (which is 179m), and the resulting wave-pattern starts at -89.5m, the point where the ship ends?
I suppose the near-field code is the one you make your money with? :)


You are definitely pushing Michlet too much if you are trying to simulate the wave pattern close to the ship. That is not what the code is for.

Good luck!
Leo.

Ah, I don't mind to refine the resulting wave-pattern. I just wanted to have indications for the reasons, thanks for your explanations.

Ciao
Niels