Michlet

[Leo Lazauskas]

Originally Posted by Martijn_vE

Daiquiry:

Thanks for the invitation. I wasn't aware of this particular problem.
Personally I haven't used the Michlet export feature anymore since I started developing DELFTship, so I hadn't noticed the problem myself.
The bug was easy enough to fix though once you know the problem.

I'm very close to releasing version 4.0, so it will be fixed with the new version.
I also increased the number of decimals for the volume to 5, which should be enough for even the smallest models.

Leo:
Wouldn't it be a good thing to display a warning/notification if the offsets are scaled, or even better prompt the user and wait for a confirmation?
I understand why you are doing this, but this can be confusing and potentially misleading, particularly to novice users.
If a user doesn't want the offsets to be scaled you could still the volume as calculated by Michlet instead of the specified volume.

Martijn:
Delftship works beautifully most of the time so I'm not surprised you didn't notice some anomalies with Michlet exports. You can get around some problems by shifting the hull in Delftship, so it just requires some care and some arcane knowledge.

My main difficulty is getting Delftship to read an offset table (in your format). I can create the table Ok from Michlet or Flotilla, but there are problems with how Delftship interprets the bowmost and sternmost sections. I think it doesn't like a section with entries that are all zeroes.

Michlet is not for novice users as I warn at the start of the manual. I also warn the user that offsets will be scaled.
7. HULL OFFSETS
It does not matter whether the offsets describing each hull are in dimensional or non-dimensional form. Michlet will automatically scale the offsets to the individual hull displacements.

I also advise users to read the manual and to work through the examples.
If they can't be bothered doing that they deserve all the consequent horrors

Cheers,
Leo.

[fredschmidt]

For all

Internet is great. This Forum is great. The peoples in this Forum are great.

But more great is to observe how the people do not have in reality frontiers.

We have here - Two Australians, a Brazilian, two Swede, a Dutch, a Italian, a Canadian working in reality for all people of the world.

Like John Lennon : For what frontiers?

[fredschmidt]

Originally Posted by Leo Lazauskas

Martijn:
Delftship works beautifully most of the time so I'm not surprised you didn't notice some anomalies with Michlet exports. You can get around some problems by shifting the hull in Delftship, so it just requires some care and some arcane knowledge.

My main difficulty is getting Delftship to read an offset table (in your format). I can create the table Ok from Michlet or Flotilla, but there are problems with how Delftship interprets the bowmost and sternmost sections. I think it doesn't like a section with entries that are all zeroes.

Michlet is not for novice users as I warn at the start of the manual. I also warn the user that offsets will be scaled.
7. HULL OFFSETS
It does not matter whether the offsets describing each hull are in dimensional or non-dimensional form. Michlet will automatically scale the offsets to the individual hull displacements.

I also advise users to read the manual and to work through the examples.
If they can't be bothered doing that they deserve all the consequent horrors

Cheers,
Leo.

Leo

You say that the offsets can be scaled. But I think, when I read, that this is an internal routine for do something.

In reality you are changing the hull permanently and this change is not more the hull from the user. By an error of input?

But the problem is exactly this. When is a error and when is not.

For any people, to no be a novice is necessary learn. But a text some times can be interpreted by different ways.

We come from the trees 14 million years ago and we have yet bigger problems in communication.

[Martijn_vE]

Originally Posted by Leo Lazauskas

My main difficulty is getting Delftship to read an offset table (in your format). I can create the table Ok from Michlet or Flotilla, but there are problems with how Delftship interprets the bowmost and sternmost sections. I think it doesn't like a section with entries that are all zeroes.

You're correct. Technically if all offsets are set to zero then to delftship that means at those locations there's no surface at all (the surface is completely on the center plane), and those panels are removed. The problem with offsets is that they are not very useful to adequately describe a hull in 3D. The fact that only half beam values can be specified at the intersection points of waterlines and stations is a serious drawback. Even worse if the stations and waterlines are spaced at constant intervals, like in Michlet. A lot of waterlines will end inbetween two stations, so they will have to be extended to the nearest stations, thus introducing distortions in the hull (and also the calculation of the volume) This also happens at stations. If the lowest point of a station falls inbetween two waterlines the station is extended to the lowest waterline.

Originally Posted by Leo Lazauskas

Michlet is not for novice users as I warn at the start of the manual. I also warn the user that offsets will be scaled.
7. HULL OFFSETS
It does not matter whether the offsets describing each hull are in dimensional or non-dimensional form. Michlet will automatically scale the offsets to the individual hull displacements.

I also advise users to read the manual and to work through the examples.
If they can't be bothered doing that they deserve all the consequent horrors

Nevertheless it's a simple fact that a lot of novice users (including professional NA's) are using Michlet due to the fact that it's available for free and relatively easy to use.
The simple fact that your warning is included in the manual (which no-one seems to read these days anyway ) doesn't present the user with the fact when it happens. If I understand you correctly the volume/offsets can either be increased or decreased depending on the outcome of the volume calculation of michlet. As I already mentioned above offsets are not adequate to accurately describe a 3D hull, so the volume calculated by Michlet will differ from the volume of the actual hull in almost any case, depending on the distortion of the waterlines/stations.
Adjusting offsets can be useful in several cases, but I also think that for many simple cases, like the one in this thread, it should be avoided or at least a warning should be displayed.


Since DELFTship seems to be used in combination with Michlet quite often it might be worth exploring other options to exchange data between the two programs, that is if you're interested.

[fredschmidt]

Martijn_vE

I do not know how are the operations that Michlet do.

But if a user see that his hull is changed how he can trust that the results can be applied for his hull?

Leo say that input volume is to use with families of hulls and now I understand why Michlet need the Volume input.

Indeed, what we really have in Michlet is a conflict of work.

The software do two works - do calculations for a family of hulls and do calculations for one specific hull, but the processing is the same for them and the conflict is here because the volume input has a precision according that it was calculated. For example, if I calculate the volume by first Simpson's rule or by polar coordinates or by trapezoidal rule or by some mathematical formula for a spline or other know curve, all be differents.

If my case is not a family of hulls and I need give a table of offsets and his volume possibly Michlet will do the correction because Michlet calculate the volume by one method and I calculate with other method and when I see the results and see that B and Cp is not of my hull I do not have confidence in results.

In reality Michelet do two differents works and I think that he need has two differents inputs.

One for family other for a single hull.

For the family volume and offsets, for single hull only offsets.

Now we have a problem. Now, I remember that when we scale a hull with the same scale for all axis the Cp do not change. But Michlet change the Cp. Cp is adimensional.

Leo, how Michlet scale the hull?


Curious

[Leo Lazauskas]

Originally Posted by fredschmidt

Leo, how Michlet scale the hull?

Michlet calculates the volume of the offsets in the table together with the user-specified length and draft. It then scales the result to the displacement that the user specifies in the input file.

Regards from sunny Adelaide,
Leo.

[fredschmidt]

Originally Posted by Leo Lazauskas

Michlet calculates the volume of the offsets in the table together with the user-specified length and draft. It then scales the result to the displacement that the user specifies in the input file.

Regards from sunny Adelaide,
Leo.

Sorry Leo. What I do not remember in that moment is that the table of offsets delivered by Freeship are wrong, are not my hull. Sorry, but for my mistake I send to you a photo of my little boat designed with Delftship, in Lake Pitangui because here is also the land of sun. Northeast of Brazil. lol

You do not like two inputs ?

[fredschmidt]

Rick

I do manual calculations (First Simpson's rule) for stations areas in a drawing obtained with Freeship for my one meter (L = 1.00 m) and I see that the results agree with data obtained from Freeship for the same hull scaled 10 .

The results are:

For hand calculations - cm²

A100 = 3
A200 = 31,27
A300 = 50,39
A400 = 62,33
A500 = 65,66
A600 = 59,03
A700 = 46,77
A800 = 28,17
A900 = 9,47
A1000 = 0

From Freeship calculations, hull scaled 10 X - m²

A1000 = 0,07
A2000 = 0,34
A3000 = 0,56
A4000 = 0.68
A5000 = 0.73
A6000 = 0.68
A7000 = 0.54
A8000 = 0.32
A9000 = 0.11
A10000 = 0

Considering errors from printing, measures, scale, process, etc I think that they agree.

What we see in the results in Hydrostatics from one meter hull, still that I do water lines spaced 1 mm apart?

When the values calculated reach .0005 the next value is rounded to .001 and we do not have values between .0005 and .00099. But if we scale 10 the hull they appear there, like you suggest.

While Freeship do not change the precision it's better work in scale 10:1, what we already do when designing RG 65 because his maximum volume is 0,001 m³

If you see the two outputs for hydrostatics together, you see this clearly.

Well, I think that for me was very profitable this thread. we understand more Freeship and Michlet and if Freeship change the precision for 6 digits ( ) and Michelet do two inputs ( ) the peace is achieved

Thanks more a time and you are a very very nice guy

Fred

[Leo Lazauskas]

Originally Posted by Martijn_vE

Adjusting offsets can be useful in several cases, but I also think that for many simple cases, like the one in this thread, it should be avoided or at least a warning should be displayed.

Since DELFTship seems to be used in combination with Michlet quite often it might be worth exploring other options to exchange data between the two programs, that is if you're interested.

I take your point about warnings, but it is not going to happen. First, my main focus with Michlet is to make it run on multiple processors which will have no screen output at all. Secondly, there is little incentive for me because it is free and I am looking for a job, so I don't have time. Thirdly, Flotilla uses a different method, i.e. it scales the offsets to the maximum beam input by the user which will confuse some users. Furthermore, Flotilla needs data above the waterline to handle cases where the hull squats at speed. Finally, Michlet is NOT a ship design program - it is unashamedly a hydrodynamics workbench (with some shippy-type stuff tacked on) so you and I have different types of users. Flotilla (my main focus at the moment) is aimed at a slightly more scientific group, i.e. those who use prefer using gnuplot and LaTeX instead of Excel and Word

I'm not sure how to make Delfthip and Michlet click together better. If Delftship could export the first (all zeroes) station and the stern (which could be all zeroes) that would make it better from my end, but it might not be practical from your end.

I accept that equally-spaced stations and waterlines are not accurate, unless many are used. I use 200 stations and waterlines (or more) when I need deadly accuracy. I think that would be far too much for Delftship to import. The latest version of Michlet also allows unequally-spaced stations as input, but ultimately the program interpolates them and ends up with an equally-spaced grid for internal computations.

Regards,
Leo.

[Leo Lazauskas]

Originally Posted by marshmat

Leo - have you decided if / when / for how much $ you'll make the Flotilla codes available? You keep teasing us about it

Demo of version 2.07 has been uploaded Mr "Pushy" Marshmat.

[daiquiri]

Leo, if I may give you a hypocritical advice (because I also use some freeware software) - don't do any charity with Flotilla. The effort you have put in it deserves a proper economic return. Your software is being used by researchers around the world, and the access to results of their work (through sciencedirect.com, springerlink.com etc.) generally requires a payment of a fee. So if they, or their academic or professional institutions, can make money by using your software, then you should make money with it as well. Good luck with your new brainchild!

[Martijn_vE]

Originally Posted by Leo Lazauskas

I take your point about warnings, but it is not going to happen.

Fair enough, it's your program.
I've fixed the problem with the zero offsets at the end, works properly now.
The only case where DELFTship removes panels during the import is if all 4 corner points of a panel are on the center plane. Zero offsets on the stern or stem are generally OK.

Of course it's possible to import 200 stations into delftship.
The software can handle this, but it will be virtually impossible to modify the hull once it has been imported.
200 stations and 10 waterlines results in approx. 2000 control points, that's a real pain

Regarding the accuracy, what is the max. amount of stations currently supported by Michlet. Still 81? You mentioned 200, is that for Michlet or Flotilla.

[Leo Lazauskas]

Originally Posted by Martijn_vE

Of course it's possible to import 200 stations into delftship.
The software can handle this, but it will be virtually impossible to modify the hull once it has been imported.
200 stations and 10 waterlines results in approx. 2000 control points, that's a real pain

Regarding the accuracy, what is the max. amount of stations currently supported by Michlet. Still 81? You mentioned 200, is that for Michlet or Flotilla.

I'd stick to 81x81 for the time being. I do supply higher resolution for people who are prepared to pay, but that's a very rare occurrence.

I would hate to try to modify a hull in Delftship that was defined with 81 stations, let alone 200!

Cheers,
Leo.

[Leo Lazauskas]

Originally Posted by daiquiri

Leo, if I may give you a hypocritical advice (because I also use some freeware software) - don't do any charity with Flotilla. The effort you have put in it deserves a proper economic return. Your software is being used by researchers around the world, and the access to results of their work (through sciencedirect.com, springerlink.com etc.) generally requires a payment of a fee. So if they, or their academic or professional institutions, can make money by using your software, then you should make money with it as well. Good luck with your new brainchild!

I refer you to a letter from Kadashman Enlil I, king of Babylon, to Amenhotep III. (See http://www.touregypt.net/amarna1.htm for full text)

Therefore, send gold, willingly, as much as you please. Send me now quickly during this summer gold in abundance, as much as is available. I could thus achieve the task I have undertaken.
(modern hieroglyphic)

Cheers,
Leo.

[baeckmo]

Originally Posted by Leo Lazauskas

Therefore, send gold, willingly, as much as you please. Send me now quickly during this summer gold in abundance, as much as is available. I could thus achieve the task I have undertaken.
(modern hieroglyphic)

Cheers,
Leo.

Amen!!!!