Flotilla 6.2 Released

Discussion in 'Software' started by Leo Lazauskas, Apr 8, 2014.

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

    Remmlinger engineer

    Here comes the explanation.
    Hopefully it is clear enough.
     

    Attached Files:

  2. 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 hope we can mesh together a bit better after I use the same Delft
    hull as you.
    (The attached graphs are for a 1.8m standard Wigley hull (L/B=10,
    L/T=16) that was used by Insel in his PhD thesis.)

    A few comments until then...

    It is interesting that you cite papers by Landweber, Patel and
    others. I have been trawling through my old BL papers from the
    Aeronautical Quarterly and elsewhere, and just re-read Granville,
    Patel's paper on thick BL, and many others by Head, Galbraith
    Perry, and others from the 1970's and 80's. I hope to find something
    reasonably fast and accurate that uses Clauser's pressure gradient
    parameter. I don't want to resort to using something slow, and
    only possibly a little more accurate, given the assumptions in thin-
    ship and linear theory already.

    I'm not sure that your assumption that the jump in the viscous drag is
    entirely due to 3D BL effects. I also get humps and hollows in the
    curve for C_F if I take into account the effect of the hull wave
    profile (HWP) for the Wigley hull. See cf.gif. The bump at about
    Fr=0.4 is largely due to the HWP. shwpons.gif shows the ratio of wetted
    area incorporating the HWP to the wet area without HWP.

    I have now included a reasonable model of the viscous pressure
    resistance into version 7.0. The difference over the simple
    viscous drag (with HWP effects) is quite small, around 2.5% to
    3.5%, as you can see in cvoponcp.gif, but it will be a little
    larger when I include the effects of the rapid BL thickening near the
    stern.

    I am a little concerned about the accuracy of the experiments
    (Delft and Wigley) at low Fr. As you can see in cwp.gif, there
    is quite a lot of scatter in the experimental data (which was
    taken from Insel's PhD thesis). At low Fr, the scatter can be as large
    as 33% although it is hard to see at the scale of the plot for the wide
    Fr range I have shown.

    I'm also not sure how close I will be able to get to your predictions
    because I'm not all that interested in building in the effects of BL
    trips. Furthermore, as you know, thin-ship theory over-estimates the
    humps and hollows at low Fr, particularly for low L/B hulls. I could
    (or might) try to smooth out those hollows by building in separation
    near the stern, but I'm not all that interested at performance for
    Fr <0.35.

    Thanks again for your interesting work!
     

    Attached Files:

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

    Remmlinger engineer

    Thanks a lot Leo for taking the time for this discussion!
    I am pretty sure about the jump in the skin friction. I can go through the program line by line in the debugging mode and for Sysser25 at FN=0.18 the B.L. is tripped at the second roughness strip, whereas at FN=0.22 transition occurs at the first strip. The strips are 0.33 meters apart with an LWL=2 meters. This occurs without any tweaking of the program or adjusting fudge-factors. The jump in the skin friction exactly mirrors the jump in the measured total drag, so I think this is causing the waviness of the total drag and it is not the humps and hollows of the wave resistance. Sysser 24 and 25 have hulls that are flat like a soap dish.

    There is a paper by Doctors and Day at the 24th Symposium on Naval Hydrodynamics 2002 about nonlinear effects. I have only a crippled version, but if I understand it right, the change in wetted surface by the HWP did not improve the results.
    I hate B.L.-trips. They are awfully difficult to model, but I had to deal with it to be able to use the Delft-results.
    Also difficult is the prediction of the wave resistance at FN<0.35. As you said the experimental scatter is large. But for a small sailing yacht, the most crucial situation, where a good VMG is essential, is beating to windward (especially if caught on a lee-shore); this happens at FN=0.35. No way to escape this regime.
    Hopefully the large number of tests in the Delft-series allows to improve the confidence by statistical means, in spite of the large scatter.
    I plan to develop a new regression analysis of the Delft-series. I am almost done, the choice of the relevant parameters is difficult.
    I'll keep you posted.
     
  4. Leo Lazauskas
    Joined: Jan 2002
    Posts: 2,696
    Likes: 155, Points: 63, Legacy Rep: 2229
    Location: Adelaide, South Australia

    Leo Lazauskas Senior Member

    Ok, I understand what you are saying. I thought it might be partly due to
    wave effects, but it is likely that it is the BL trips causing the large changes.

    I just emailed you a collection of papers by Lawry Doctors, including the full
    paper from the 24th Symposium.
    It wasn't the changes in HWP that Doctors and Day showed had little effect
    on C_F, but their "enhanced" viscous drag calculation. That enhancement
    was just a simple addition of the BL to the offsets and didn't include the HWP.
    The reason their enhancement had little effect was because it induced a
    change in the squat. The additional drag causes the hull to rise slightly more
    which in turn reduces the effective wetted surface area. The nett effect is
    very small, as can be seen in their paper. The way that the viscous drag
    enters into squat calculations is given in their paper in equation (12) on
    page 6.

    I put that calculation into Flotilla 7 and it indeed makes only a small
    difference. It's in the code now and takes very little time to calculate so
    I'll leave it in.

    If you have the time, please send me the offsets of a Delft hull that you
    think is a good representative, and I'll use it for my new calculations.
    I'm happy to keep posting results here, or we can start a new thread in the
    Hydrodynamics forum.
     
  5. Leo Lazauskas
    Joined: Jan 2002
    Posts: 2,696
    Likes: 155, Points: 63, Legacy Rep: 2229
    Location: Adelaide, South Australia

    Leo Lazauskas Senior Member

    Uli:
    I re-read my post and noticed that I was unclear about which component
    Doctors and Day showed wasn't affected much by their "enhanced viscous"
    drag method.

    Their Figure 8 shows the effect on the pressure resistance component
    R_P. The effect of adding the boundary layer displacement is indeed
    tiny.

    They also investigated how the HWP affects resistance, but they did so
    in a non-standard way.
    Their NL-2 and NL-3 models moved the hull offsets up and down along the
    length of the hull according to the HWP at each station. The hull was
    distorted in such a way that the free-surface was flat at the
    waterplane. They then calculated the total resistance and squat for
    that distorted vessel. These two models gave very poor predictions,
    which is not surprising because the free-surface boundary condition is
    incorrect. Simply put, the "linearized" free-surface boundary
    condition is not the same as a flat free-surface along the length of
    the hull.
     
  6. Remmlinger
    Joined: Jan 2011
    Posts: 312
    Likes: 58, Points: 28, Legacy Rep: 172
    Location: Germany

    Remmlinger engineer

    Thanks for this eye-opener, this is key in understanding why they failed!

    Nevertheless I think adding nonlinear effects is often not worth the effort.
    Some time ago I used the HWP from experiments to adjust the wetted girth-length of the sections. The skin friction drag, that I get as a result of the integration of the shear-forces along the hull, changed very little and the comparison with the measured drag did not improve. If I used this adjusted girth-length also in the subroutine that computes the potential flow, the program did not converge any more.

    In my correlation I lump these effects all together into the residuary resistance.
    Since you calculate the true wave resistance, you have to deal with all these effects, at least if they have a significant influence - not an easy task. ;)
     
  7. Leo Lazauskas
    Joined: Jan 2002
    Posts: 2,696
    Likes: 155, Points: 63, Legacy Rep: 2229
    Location: Adelaide, South Australia

    Leo Lazauskas Senior Member

    1. It was a reasonable method to try. If successful, it would have meant
    that some non-linear effects could have been accounted for relatively
    quickly. I tried something similar at the same time and my results
    were so embarrassingly bad that E.O. Tuck didn't want to publish.
    I'm glad Doctors and Day published: it's sometimes important to
    show dead ends to others so they don't waste time themselves. :)


    2. I believe it is worthwhile for some hulls to include the HWP.
    I have only come to understand this and how it relates to squat after
    working with some experiments of hulls that were held fixed and not
    allowed to sink and trim. In those cases, the water level drops around
    the hull which affects both the viscous drag and wave resistance.
    When allowed to squat, the hull settles down into the hollow
    surrounding the hull. This doesn't affect the wave resistance much. As
    Prof. S. D. Sharma put it in his comments on another paper by Doctors
    and Day:
    "The reason why the use of measured trim and sinkage in calculating the
    flow by Michell's theory is counterproductive may be that in model
    experiments the sinkage is measured relative to the carriage, not
    relative to the surrounding water surface. In reality, most of the so-
    called sinkage is simply due to the mean local lowering of the free
    water surface around the ship. It does not really affect the wave-
    making potency of the hull."

    It does, however, affect the skin-friction. I tried several different
    methods previously (in Michlet and elsewhere), but the effects were
    only marginally successful. I have now implemented a much better method
    in the latest version of Flotilla.

    3. In earlier versions of Flotilla I lumped several effects together in
    the "transom stern hollow" resistance. This component included the
    loss of hydrostatic pressure on the dry portion of the transom, as well
    as BL effects and the wave-making of the transom. It was not effective
    for hulls without transoms, however. I am working on that now.

    4. That task is not as difficult as it may seem because I already
    calculate the pressure on the hull to estimate squat.


    5. In the report you attached in your earlier post, the skin-friction
    coefficient goes to zero at the stern.
    Is that because the hull wetted area is zero at the stern (i.e. is it
    pointed for the Delft hull) or is it because the shear stress is zero
    because the BL is separating at or near to the stern?

    In some work a few years ago, I implemented Stratford's turbulent
    separation criterion, but it made very little difference because it
    predicted separation very close to the stern, at least on well-
    streamlined hulls. I then used Stratford's laminar separation criterion
    which had a greater effect. I guess I could justify that criterion on
    the basis that laminar separation is likely to precede turbulent
    separation, but I'm not completely sure of that, especially when there
    is a free-surface involved.
     
  8. Ad Hoc
    Joined: Oct 2008
    Posts: 7,773
    Likes: 1,678, Points: 113, Legacy Rep: 2488
    Location: Japan

    Ad Hoc Naval Architect

    Oh well said. Research is not just about what does work...
     
  9. Remmlinger
    Joined: Jan 2011
    Posts: 312
    Likes: 58, Points: 28, Legacy Rep: 172
    Location: Germany

    Remmlinger engineer

    The local shear stress is zero because of separation. In addition the wetted area approaches zero at the stern, so the local shear force is zero times zero. Therefore the accuracy of the BL-simulation is not so important at the stern and the integral of the shear forces along the hull yields very robust results. This is much in contrary to the methods of Squire-Young and Granville, who rely on the BL-thickness at the trailing edge. This thickness in the vicinity of the stern is hard to predict (as Patel explained).
     
  10. Leo Lazauskas
    Joined: Jan 2002
    Posts: 2,696
    Likes: 155, Points: 63, Legacy Rep: 2229
    Location: Adelaide, South Australia

    Leo Lazauskas Senior Member

    Thanks, AdHoc.
    If I had got even slightly better, converged results I might have published.

    Funding bodies don't like to see words like "failure" in publications unless it is aimed at your competitors. :)
    The worst transgressors are in the pharmaceutical industry/racket. It is a real
    shame that they don't publish negative or marginal experimental studies.
     
  11. 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 suspect that one reason that I get reasonable predictions for transom stern
    vessels is that the longitudinal slope at the stern is smaller than for cruiser
    or pointed sterns, and so the BL doesn't thicken so quickly. Therefore, like
    your model, the modelling is not so sensitive near the stern, especially if
    separation occurs at the exact trailing edge.
     
  12. Ad Hoc
    Joined: Oct 2008
    Posts: 7,773
    Likes: 1,678, Points: 113, Legacy Rep: 2488
    Location: Japan

    Ad Hoc Naval Architect

    We have presented some of our finest failures ..much to the surprise of many :)

    But if one does not learn from the mistake, where is the progress...and how does one educate the wider community to prevent others making the same? It is our responsibility to educate the next generation..by whatever means...
     
  13. johnhazel
    Joined: Jun 2008
    Posts: 250
    Likes: 5, Points: 18, Legacy Rep: 60
    Location: Michigan

    johnhazel Senior Member

    Flotilla 6.2 running "monoexwp" on i7-4790k at stock speed:

    0.23431E+2 seconds (wall clock)
     
  14. Leo Lazauskas
    Joined: Jan 2002
    Posts: 2,696
    Likes: 155, Points: 63, Legacy Rep: 2229
    Location: Adelaide, South Australia

    Leo Lazauskas Senior Member

    Thanks John.
    I have included a fair bit more for the next release, but I'm in the middle of finishing of a rowing program I hope to release within the month.
    That is taking a while to document because there are so many things to specify, e.g. limb lengths and their cg's, oar details, forces, rigging, body angles as functions of time etc. It's not too bad for one-man boats, but a bit of a nightmare for eights!
     
    1 person likes this.

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

    Remmlinger engineer

    Christmas present

    For all those people who are looking for catamaran tank-test data to validate their calculations, there is a weblink that might not be known:
    The publication is described on this page: http://canal.etsin.upm.es/publicaciones/articulos.html
    The weblink is: http://canal.etsin.upm.es/papers/catamaran/
    There are excel-files with the resistance and the wave elevation and even an IGES-file for the 3-D hullform. It is a hard chine model.
    The first 2 pages of the paper are available at Springer, see attachment.
    Uli
     

    Attached Files:

    1 person likes this.
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.