Godzilla; hard chines and other questions.

[Tcubed]

I'm getting to grips with michlet/godzilla, and have read the manuals as well as searching this forum for insight and tips, but i still have some difficulties.

Specifically, what is the constraint that i need to use in order to keep all sections made up of straight lines , ie; hard chines? And if that is even possible, how to specify the number of chines?

I overstand that it will optimize to whatever the constraints are, thus am i correct in thinking that the starting hull shape is irrelevant? Of course if the starting hull shape is very different to what is optimal for the requirements (specifically speed and displacement) it will take longer to converge, but the end result should be the same...?

Mr Willoughby uses ;
# ============================ FIRST HULL ==============================
# Offsets
1, 1.0, 1, 0.0

Why those numbers specifically?
And what do those numbers represent?

Am i correct in overstanding that the ranges specified in = SHIP CONSTRAINTS = must be equal to or greater than those specified in = First HULL CONSTRAINTS = ?
Of course the FIRST HULL values must be within the smallest ranges specified.

For assymetry; the tendency seems to be to be fairly consistently towards high levels of for and aft symmetry. This though begs the question of how squat is accounted for.. For instance at Froude #s a bit over 0.4 most hulls will tend to squat, even if long skinny hulls tend to squat fairly subtly.
It also makes me wonder how michlet deals with the offsets in this case since it only considers the underwater shape as per hydrostatic flat water - however, underway portions of the boat that were dry when stationary become immersed due to bow and stern wave as well as squat.

How do i specify the last shape function triplet when using hull series 8? What does it represent to do with the transom exactly?

I include the .fbm file for an efficient hull i am designing in delftship. Stability requirements are nil as it will be hydrofoil (+one small ama for when stationary) stabilized.
As can be seen i stuck to fairly traditional shape concepts; shallow deadrise aft to minimize squat, steep deadrise and fine entry for ease in chop, moderate Wlbeam/draft ratio to minimize wetted area, minimal curvature aft in the buttocks , etc.
My optimization requirements are minimum resistance at high speeds - in the 20 to 25 knot range - i am not too concerned about optimization at lower cruising speeds like 7 to 12 knots as the resistance in that range will be much lower anyways.

I exported this to michlet specifying 1 to 25 knot speed range and get the screenshot included. At 25v knots (12.861 m/s) i get 1.052 kN resistance.

The latest test i did in godzilla is included using series 8 hull
Must say i am a little surprised at the amount of immersed transom given as well at how short it ends up.
I wonder if accuracy in michlet starts to break down when there is so much immersed transom.

With this last run the file fsoff1.txt just says # BLANK
I want to import>surface to delftship
Is that a problem with the series 8 or is it that i have to wait till Godzilla finishes? It is already at 137 000 Evals and the hull does not seem to be hardly changing anymore and the population is all very similar so can i tell it i'm already happy with that shape , or do i have to wait till the end?
Is there a way to specify how many evaluations i want done?

I know this is a lot of questions but so far i'm reading the manuals and iterating runs changing just one number aty a time and it takes a long time, so any pointers will be very much appreciated as they will shorten the learning curve very considerably.

[Tcubed]

I see that it does eventually stop after 10^6 Evals.
but the fsoff1.txt file is stil empty, which is a shame as i wanted to have a good look at it in one of my modeling programs or delftship.

It ends up with 0.9766 kN resistance.

On another note; does anyone use flow-3D?

[Leo Lazauskas]

Originally Posted by Tcubed

Specifically, what is the constraint that i need to use in order to keep all sections made up of straight lines , ie; hard chines? And if that is even possible, how to specify the number of chines?

No, that is not possible in the public release of Michlet.

Take heed of what it says on the first page of the manual. Michlet is NOT a ship design program. It is a small hydrodynamics workbench.

Originally Posted by Tcubed

I overstand that it will optimize to whatever the constraints are, thus am i correct in thinking that the starting hull shape is irrelevant? Of course if the starting hull shape is very different to what is optimal for the requirements (specifically speed and displacement) it will take longer to converge, but the end result should be the same...?

Mr Willoughby uses ;
# ============================ FIRST HULL ==============================
# Offsets
1, 1.0, 1, 0.0

Why those numbers specifically?
And what do those numbers represent?

The first hull is just a dummy. It doesn't affect the optimisation.

The method will often come up with similar hulls, but it might also find a different shape that has similar resistance. It is then up to you to assess those competing candidate vessels during your tour around the "design spiral".

The shape parameters are described in the manual.

Originally Posted by Tcubed

Am i correct in overstanding that the ranges specified in = SHIP CONSTRAINTS = must be equal to or greater than those specified in = First HULL CONSTRAINTS = ?
Of course the FIRST HULL values must be within the smallest ranges specified.

The principal parameters describing the first hull should lie between the ship constraints.

Originally Posted by Tcubed

For assymetry; the tendency seems to be to be fairly consistently towards high levels of for and aft symmetry. This though begs the question of how squat is accounted for.. For instance at Froude #s a bit over 0.4 most hulls will tend to squat, even if long skinny hulls tend to squat fairly subtly.
It also makes me wonder how michlet deals with the offsets in this case since it only considers the underwater shape as per hydrostatic flat water - however, underway portions of the boat that were dry when stationary become immersed due to bow and stern wave as well as squat.

In Michell's thin-ship theory it can be shown that wave resistance is a minimum for fore-aft symmetric hulls.

Squat is definitely not included in the calculations. For that you need to use other programs.

Originally Posted by Tcubed

How do i specify the last shape function triplet when using hull series 8? What does it represent to do with the transom exactly?

The latest test i did in godzilla is included using series 8 hull
Must say i am a little surprised at the amount of immersed transom given as well at how short it ends up.
I wonder if accuracy in michlet starts to break down when there is so much immersed transom.

The manual describes the different simple shape functions.

Large transom sterns violate the small longitudinal slope assumption in thin-ship theory.

Originally Posted by Tcubed

With this last run the file fsoff1.txt just says # BLANK
I want to import>surface to delftship
Is that a problem with the series 8 or is it that i have to wait till Godzilla finishes? It is already at 137 000 Evals and the hull does not seem to be hardly changing anymore and the population is all very similar so can i tell it i'm already happy with that shape , or do i have to wait till the end?
Is there a way to specify how many evaluations i want done?

I'm not sure if Delftship still accepts Michlet output.

You have to wait until Godzilla finishes, or stop it, and then calculate the drag before exiting.

You can stop Godzilla at any time you like, calculate the drag or wave pattern and exit. You cannot specify the number of evaluations in the public release.

Good luck!
Leo.

[Tcubed]

Many thanks for that reply Leo. Would have responded sooner but have been very busy building myself a dinghy, working on my sailboat, etc.

Ok no parameters exist in order to restrict the optimization to hard chine sections.

Now i see also that one must be quite careful interpreting the results since squat and dynamic lift are not considered. So if it optimizes to a shape that would entail those factors considerably it would be safer to be skeptical of the results.
However , the results can be taken top be exceptionally accurate for hull shapes which do not have squating and/or planing tendencies (such as typical multihull amas and rowing shells).

However i do keep on getting a blank fsoff1.txt file...
I did get a valid fsoff1.txt file in one my early runs but i can't remember what i did different.
This is really stumping me because you can indeed import this text file to delftship (and hence on to rhino, etc) by clicking import>surface and browsing for the fsoff1.txt file. This took me awhile to figure out but i was very happy to figure it out as it opens up huge possibilities.

Any ideas why i might be getting a blank fsoff.txt file?

[Leo Lazauskas]

Originally Posted by Tcubed

Now i see also that one must be quite careful interpreting the results since squat and dynamic lift are not considered. So if it optimizes to a shape that would entail those factors considerably it would be safer to be skeptical of the results.
However , the results can be taken top be exceptionally accurate for hull shapes which do not have squating and/or planing tendencies (such as typical multihull amas and rowing shells).

I wouldn't go quite so far as saying it is exceptionally accurate.
The thin-ship approximation of the wave resistance gets better the thinner the hull. You still have skin-friction, separation, roughness and transom stern effects, among others, to worry about.

The effect of squat on wave resistance is complicated by what Don MacPherson called "the moving dish problem". As a hull squats, the surface of water also changes. If the hull settles into the "dish" then its wetted area doesn't change much from the static value, and the hull's wave-making potency is similar to the static attitude. (Interestingly, for hovercraft the wave resistance can actually be estimated from the value of the trim of the vessel).

In thin-ship theory, the effect of squat should be small. If it is not, then we should probably also include other second-order effects to be consistent.

Another program I wrote, "Flotilla", calculates sinkage, trim, and dynamic forces and moments. I will try to release an updated version here when I get time. (I first want to finish the Surface Effect Ship version before re-visiting the other versions).

Originally Posted by Tcubed

However i do keep on getting a blank fsoff1.txt file...

I'll have a look when I get some time.
Which version are you using?

Leo.

[Ad Hoc]

Originally Posted by Tcubed

However , the results can be taken top be exceptionally accurate for hull shapes which do not have squating and/or planing tendencies (such as typical multihull amas and rowing shells).

Be carefull. All boats trim and squat no matter their shape/size. The influences on the magnitude of the trim/squat are: length, displacement, speed and hull shape plays a part but that is a consequence of the aforementioned.

[Tcubed]

I'm not too sure which version i have. I downloaded it a couple of years ago from somewhere on this forum but only lately managed to get around to figuring out the GODZILLA part.

The executable says mlt807.exe.

I got the spanish tutorial also.
I've read through the included html help files as well, although no doubt it would be worth re- reading as i may have missed a few things. In particular i'm still a bit unsure as to exactly in what way the parameters restrain the shape.

[DMacPherson]

Just a bit of clarification as it pertains to Leo's quote of my comments about the "moving dish". (First, I did not quoin the phrase. Don't remember where it came from, but it may have been from Saunders.) The "moving dish" surface depression is principally an effect of motion through shallow water. In open water this effect is small and relatively inconsequential.

Thus, I think of squat as a shallow water artifact, and sinkage as the typical dynamic vertical change in CG that occurs as a hull develops a lower pressure distribution under the hull. The center of the low pressure is relative to the CG is the cause of the trimming. Most fast ships will trim by the stern, but large slow ships in very shallow water can trim by the bow.

The "moving dish" become important when considering under-keel clearance, as the loss of clearance is the sum of the dynamic sinkage plus the lowering of the effective water surface.

Don MacPherson

[Leo Lazauskas]

Originally Posted by DMacPherson

Just a bit of clarification as it pertains to Leo's quote of my comments about the "moving dish". (First, I did not quoin the phrase. Don't remember where it came from, but it may have been from Saunders.) The "moving dish" surface depression is principally an effect of motion through shallow water. In open water this effect is small and relatively inconsequential.

Thus, I think of squat as a shallow water artifact, and sinkage as the typical dynamic vertical change in CG that occurs as a hull develops a lower pressure distribution under the hull. The center of the low pressure is relative to the CG is the cause of the trimming. Most fast ships will trim by the stern, but large slow ships in very shallow water can trim by the bow.

The "moving dish" become important when considering under-keel clearance, as the loss of clearance is the sum of the dynamic sinkage plus the lowering of the effective water surface.

Don MacPherson

Hi Don,
you were the last person to use that elegant phrase on boatdesign,net, so I attributed it you for the purposes of the last post.

The "moving dish" also occurs in deep water, but for thin hulls it isn't all that important, except perhaps for transom dryness at lowish Froude numbers.

It can also come into play on some hulls of small draft such as rowing shells, surfboards, and other slender bodies, especially if they are not moving steadily - the water level rises and falls and this can increase and decrease the wetted area and the wave-making potency of the hull.

I hope all went well with your Young N&Me Papers Day!
Leo.

[Leo Lazauskas]

Originally Posted by Tcubed

I'm not too sure which version i have. I downloaded it a couple of years ago from somewhere on this forum but only lately managed to get around to figuring out the GODZILLA part.

The executable says mlt807.exe.

I got the spanish tutorial also.
I've read through the included html help files as well, although no doubt it would be worth re- reading as i may have missed a few things. In particular i'm still a bit unsure as to exactly in what way the parameters restrain the shape.

For the three parameter series, the 1st parameter controls the waterplane shape, the 2nd controls the cross-section shape, and the 3rd controls the keel-line.

A value of 0.0 means the shape is rectangular, a value of 0.5 means elliptical, and a value of 1.0 means parabolic.

Thus, 0.0, 0.0, 0.0 is a rectangular block.

1.0, 0.0, 0.0 is a (vertical) parabolic strut, i.e. it has parabolic waterplanes, rectangular cross-sections, and a rectangular side-view.

0.5, 0.5, 0.5 is a spheroid. If the length and beam are identical, and the draft is half the length, it is a hemisphere.

1.0, 1.0, 0.0 is a Wigley parabolic hull.

1.0, 0.5, 1.0 is a fine canoe body with parabolic waterlines, elliptical cross-sections and a parabolic side-view.

You can also use other values: thus, 0.75 0.0, 0.0 is a vertical strut with waterlines that are between an ellipse and a parabola.

I devised this particular series so I had a convenient, intuitive (for me) way of easily imagining the hull shapes. If you give me three parameters for a hull, I can immediately picture it in my mind without having to draw it.

Many other hull series in Michlet (e.g. Series 7 etc) just build on that basic 3 parameter structure. Again, with experience, I can imagine what a 7 parameter hull is like without having to draw it.

Good luck!
Leo.

[Tcubed]

Ok Thanks a million .
That really clears it up well.
I'm going to play around with those parameters now that you've given me a very clear concept as to how they work.

[DCockey]

Originally Posted by DMacPherson

Just a bit of clarification as it pertains to Leo's quote of my comments about the "moving dish". (First, I did not quoin the phrase. Don't remember where it came from, but it may have been from Saunders.) The "moving dish" surface depression is principally an effect of motion through shallow water. In open water this effect is small and relatively inconsequential.

Thus, I think of squat as a shallow water artifact, and sinkage as the typical dynamic vertical change in CG that occurs as a hull develops a lower pressure distribution under the hull. The center of the low pressure is relative to the CG is the cause of the trimming. Most fast ships will trim by the stern, but large slow ships in very shallow water can trim by the bow.

The "moving dish" become important when considering under-keel clearance, as the loss of clearance is the sum of the dynamic sinkage plus the lowering of the effective water surface.

Don MacPherson

I'm confused. I don't see a fundamental difference between "squat" and "sinkage" unless squat is a change of trim angle and sinkage is a change of CG height relative to the undisturbed free surface. Generally boats will do both. The same physical phenomena cause both change of trim angle and height relative to free surface, with the magnitudes varying depending on the depth of the water.

[Leo Lazauskas]

Originally Posted by DCockey

I'm confused. I don't see a fundamental difference between "squat" and "sinkage" unless squat is a change of trim angle and sinkage is a change of CG height relative to the undisturbed free surface. Generally boats will do both. The same physical phenomena cause both change of trim angle and height relative to free surface, with the magnitudes varying depending on the depth of the water.

For me, squat means sinkage and trim induced by a hydrodynamic forces and moments. At high speed the sinkage force reverses direction and the hull starts to have some of its weight supported by the forces induced by the passage of water around the hull. At still higher speeds, most of the weight of the hull is supported, and I call that "planing".

I prefer this "continuous" definition of squat rather than ones that are limited to finite depth water and to downward forces only.

Leo.