Flotilla 6.1 Released

Discussion in 'Software' started by Leo Lazauskas, Feb 27, 2014.

  1. Leo Lazauskas
    Joined: Jan 2002
    Posts: 2,696
    Likes: 155, Points: 63, Legacy Rep: 2229
    Location: Adelaide, South Australia

    Leo Lazauskas Senior Member

    The gnuplot page is:
    http://www.gnuplot.info/

    There are pre-compiled versions available from:
    http://www.tatsuromatsuoka.com/gnuplot/Eng/winbin/

    Download by clicking on the one that starts with...
    0001 gp470-20120916-win32-mingw-setup.zip...

    I have put up the links to Gnuplot (and, WinEdt, a good text editor) on the Flotilla
    page so others don't have the same problems.)
     
    Last edited: Mar 1, 2014
  2. Leo Lazauskas
    Joined: Jan 2002
    Posts: 2,696
    Likes: 155, Points: 63, Legacy Rep: 2229
    Location: Adelaide, South Australia

    Leo Lazauskas Senior Member

    A tip for those investigating vessels with transom sterns...

    The only indication to Flotilla (and Michlet) that a hull has a
    transom stern is if the offsets (table or spline) has some non-zero
    values in the last row.

    Some hull drawing programs do not always produce a valid set of
    offsets so you should always inspect their output before using them
    in Flotilla.

    A common problem is that those programs will output a table of offsets
    with all zeroes in the last row. This often happens if the hull has a
    slightly curved transom when viewed from above. The drawing programs
    sometimes put the last station at where the curve is a maximum, and
    which therefore has zero thickness. In this case, it is better to put
    the last station a little forward of the exact end of the hull.

    If the offsets do not have non-zero offsets when they should, the wave
    resistance will be very high because Flotilla will assume a cruiser stern with
    a large exit angle.
     
    Last edited: Mar 3, 2014
  3. Leo Lazauskas
    Joined: Jan 2002
    Posts: 2,696
    Likes: 155, Points: 63, Legacy Rep: 2229
    Location: Adelaide, South Australia

    Leo Lazauskas Senior Member

    There is a minor bug in Flotilla 6.1.

    The Beam for the submerged portion of the hull is shown on-screen as the
    overall beam, not the BWL. The correct values are saved to the geo.gpl file,
    just not reported on-screen.

    I have fixed the problem and added a few small extra quantities such as the
    (longitudinal and transverse) metacentric heights.

    Should I release the corrected exe file here? Or in a new thread?
    Or should I wait and release 6.2 with a few more features?
    I deliberately left out a lot of output and other detail in the Demo Version, not
    to be mean, but because the program is already very complicated. I will slowly
    add these extra features in up-coming versions.
     
  4. DCockey
    Joined: Oct 2009
    Posts: 5,229
    Likes: 634, Points: 113, Legacy Rep: 1485
    Location: Midcoast Maine

    DCockey Senior Member

    Leo, would you recommend Michlet or Flotilla for analyzing hull shapes of a double ended boat with 6:1 length/beam ratio, shallow draft (small fraction of beam), Cp of 0.45 to 0.6, and Fn of 0.25 to 0.33?
     
  5. Leo Lazauskas
    Joined: Jan 2002
    Posts: 2,696
    Likes: 155, Points: 63, Legacy Rep: 2229
    Location: Adelaide, South Australia

    Leo Lazauskas Senior Member

    It would probably be pushing both programs a bit, David.

    Firstly, it is likely that they will both exaggerate the humps and hollows in
    the wave resistance curve. That wouldn't be as much of a problem for a
    transom stern vessel with similar proportions because of the smaller
    longitudinal slope I mentioned in an earlier post. I have another code that
    reduces the humps and hollows (using a surface viscous damping formulation
    due to Lamb) but that is nowhere near ready for routine use.

    Secondly, at those small Fn viscous drag is likely to be very important, so
    you will need to come up with a form factor of some sort. The form factor
    in Flotilla might help, but it also might be a bit low. If you can get some
    guidance from other work on similar hulls that would be very useful.

    Thirdly, the shallow draft could be a problem. Beam-to-draft ratios of 4 or so
    should be Ok; B/T > 6 or more could be a problem.

    If you have set up an input file for Michlet it is relatively straight forward to
    convert it to Flotilla format. I am absolutely certain of this because I'd do it
    for you, gratis, in less than 10 minutes. :)

    Leo.
     
  6. Remmlinger
    Joined: Jan 2011
    Posts: 312
    Likes: 58, Points: 28, Legacy Rep: 172
    Location: Germany

    Remmlinger engineer

    Thanks to the rainy weekend I got the time to play a little with flotilla.
    It seems to me, that the computed drag of slender ships without a transom stern is the same in Michlet and Flotilla. The results for squat are an additional bonus in Flotilla, but the price is high. Michlet is done within a second, whereas Flotilla ran for 64 minutes. The model used was the Inui S-201 hull and 81 speeds between Fn=0.1 and Fn=0.8. Result is attached.
    Changes of the viscous damping (nut) had no effect on the computed drag.
    Leo, is this in line with your intentions?
    This is no critique, I still think your programs are a great gift. The prediction of the resistance for the Inui-hull above Fn=0.35 is marvellous.
    Uli
     

    Attached Files:

  7. Leo Lazauskas
    Joined: Jan 2002
    Posts: 2,696
    Likes: 155, Points: 63, Legacy Rep: 2229
    Location: Adelaide, South Australia

    Leo Lazauskas Senior Member

    Thanks, Uli.

    I haven't tried comparisons with the Inuid hulls myself, but your
    predictions are quite good. I'm not surprised by the poor agreement
    at low Fr because the Inuid has a rounded stem and stern which would
    violate the thin-ship assumption of small longitudinal slope in their
    immediate vicinity. Once the wavelengths become longer with increasing
    Fr the predictions settle down to something reasonable.

    And, yes, the price for squat calculations is quite high. I wish more
    CFD publications gave wall-clock times and details of the computers
    required to run them. I suspect that calculating squat and resistance
    at 81 speeds could take a month or more on a single processor of a PC,
    perhaps longer than 3 months for some URANS codes like CFDShip.

    You are correct that the turbulent eddy viscosity "nut" has no effect
    on the resistance of hulls without transoms. Viscous wave damping only
    affects wave elevations. For a transom stern hull there are some weak
    Reynolds number effects on the wave height at the edge of the transom.
    The BL displacement thickness around the edge of the transom also has
    a minor effect on resistance, especially on very small model hulls.
    These do not have any effect on pointed or cruiser sterns.

    I will release an update of Flotilla in the next week or so with some
    minor bug fixes and some additional output. Until then, note...

    * Flotilla needs at least 2 speeds in the udrag.flo file if you are
    specifying individual speeds. If you put in only one speed it won't run
    and it doesn't produce an error message.

    * The on-screen value of the beam for the submerged portion of the hull
    was the overall beam (BOA), not the beam at the waterline (BWL). This
    glitch had no effect elsewhere: the correct value of BWL was used in
    all calculations.

    * Although Flotilla is not a hull design program, some users wanted a
    few additional calculations such as the LCB, VCB, LCF and longitudinal
    and transverse metacentric heights that they could use as constraints
    during optimisation runs. These quantities will be output in the next
    version.

    * Wave elevations very close to the hull (e.g. within one beam) are now
    more accurate.

    * More examples, e.g. Delft 372 demihulls. (I will release the
    catamaran and SES versions when I get some time.)

    * Very narrow deep transoms on small hulls will produce unrealistically
    high "hollow resistance". Until the release, be aware of this and don't
    include the hollow resistance in the total if it looks too high, e.g.
    10 or more times the wave resistance. If you are modelling a cruiser
    stern, then make sure the offsets are all zero at the stern.

    Leo.
     
  8. Remmlinger
    Joined: Jan 2011
    Posts: 312
    Likes: 58, Points: 28, Legacy Rep: 172
    Location: Germany

    Remmlinger engineer

    Leo, though I am not one of the attractive Russian girls that you are used to (rather a grey weirdo), I would still like to ask your opinion.

    The Inui-hull has a half entrance angle of 21 degrees (lines plan attached). Would you really consider this a violation of the thin-ship assumption?
    The wavy curvature between the stations is a flaw of Delftship. In Rhino the curvature is smooth.

    I am interested in the drag-difference that is caused by a small change of the Inui-hull (e.g. a slight change of the keel profile). I have calculated the drag for both hull varieties with Michlet, calculated the small difference and applied this difference to the measured drag of the original hull. Would this be an acceptable estimate for the drag of the tweaked hull in the opinion of a mathematician? I know it is more dirty than quick, but I have no better idea.
     

    Attached Files:

  9. Leo Lazauskas
    Joined: Jan 2002
    Posts: 2,696
    Likes: 155, Points: 63, Legacy Rep: 2229
    Location: Adelaide, South Australia

    Leo Lazauskas Senior Member

    The angles at the bow and stern are a bit wide for a thin-ship method,
    but reasonable if you want to compare the performance of two similar
    designs, for example by tweaking the keel-line of one hull.

    So, yes, I think your idea is a perfectly acceptable engineering method.
    If the adjustments are reasonably small I doubt whether you could
    discern the differences with tank tests, unless they were performed very
    accurately and run several times at the same speed. That was also the
    conclusion of Scragg and Nelson who wrote a very good paper, "The Design
    of an Eight-Oared Rowing Shell". They couldn't find any sensible trends
    because of the scatter in their experimental data, but they were pleased
    with the trends predicted by thin-ship theory.

    Leo "another grey weirdo" Lazauskas
     
  10. johnhazel
    Joined: Jun 2008
    Posts: 250
    Likes: 5, Points: 18, Legacy Rep: 60
    Location: Michigan

    johnhazel Senior Member

    0.51780E+02 seconds (wall clock)

    i3-2310M 2.1Ghz 4Gb Windows 7 Home Premium

    Windows experience index Processor: 6.4, Memory 7.3
     
    1 person likes this.
  11. Remmlinger
    Joined: Jan 2011
    Posts: 312
    Likes: 58, Points: 28, Legacy Rep: 172
    Location: Germany

    Remmlinger engineer

    Comparison Michlet-Inuids

    I have reworked Inui's original test results and compared them to Michlet.
    Diagrams and numerical values are displayed here:
    http://www.remmlinger.com/Inui S-201.pdf
    The closely spaced residual resistance values might be useful for other validation work too.
    Uli
     

  12. Remmlinger
    Joined: Jan 2011
    Posts: 312
    Likes: 58, Points: 28, Legacy Rep: 172
    Location: Germany

    Remmlinger engineer

    For those who downloaded the report (quite a few according to my website statistics): please do it again, because there is a small correction now that affects the numerical values of the viscous coefficients.
     
Loading...
Forum posts represent the experience, opinion, and view of individual users. Boat Design Net does not necessarily endorse nor share the view of each individual post.
When making potentially dangerous or financial decisions, always employ and consult appropriate professionals. Your circumstances or experience may be different.