Michlet

Discussion in 'Software' started by pavel915, Dec 8, 2009.

  1. Guest625101138

    Guest625101138 Previous Member

    The Wigley hull does not have a transom!

    You keep on about a BIG discrepancy. How was the wave drag isolated from the test data? What is the big discrepancy in overall drag of the hull? What depth was the water? What temperature was the water? Was it salt water or fresh? Did they allow the model to sink/lift and trim?

    Look at the attached chart of total resistance for a Wigley hull. Lets take the worse case around Fn 0.5. The Michlet Ct is 8.7E-3 the test data is 7.9E-3. So an overestimate of 10%. I have no idea how many of the possible variables have been aligned between the model and the testing but there is plenty of room for error.

    I use Michlet as the basis for building boats that are of a usable scale and validate the Michlet data accordingly rather than using little models. It would be nice to have Flotilla to cater for more factors but these are quite small when you are aiming to build efficient hull forms and I can live with Michlet particularly when it is paired with GODZILLA.

    Rick W
     

    Attached Files:

  2. Guest625101138

    Guest625101138 Previous Member

    I am reasonably confident that they have zero hydrostatic pressure because they have attached flow off the transom. The hull actually has a very bluff canoe stern rather than a clean cut.

    I have seen this error before. See my screen dumps above. The hydrostatic pressure has a slower rise with speed as evidenced when the in.mlt file is correct. There is no hydrostatic pressure in the case where the in.mlt file is faulty indicating that Michlet is not seeing the transom.

    Rick W
     
  3. daiquiri
    Joined: May 2004
    Posts: 5,371
    Likes: 258, Points: 93, Legacy Rep: 3380
    Location: Italy (Garda Lake) and Croatia (Istria)

    daiquiri Engineering and Design

    Correct. My fault, I'm posting without re-reading before. So not only wet transom issues.
    I have no idea and don't want to guess. Hopefully they will reply and tell us how was the wave drag isolated from the total. We will add this question to the list.
    Where did you take this one from? This discussion is based on Formsys' official validation page and their graphs.

    I must admit it's funny that you are fighting for the Michlet cause more fiercely then Leo does...
     
  4. Guest625101138

    Guest625101138 Previous Member

    Yes lots of room for error. My test data is much closer to Michlet than 10% when factors such as temperature, depth and water density are aligned. Flotilla closes the gap even more when sink/lift and trim are accounted for.

    Apart from Olympic rowers, Olympic paddlers and Marathon Pedallers, who is aiming for better than 10% on a bare hull prediction for real world engineering? Throw in a bit of surface fouling, an appendage or two, windage, some wind waves and so on and, from an engineering perspective, 10% is outstanding. Being conservative rather than optimistic is on the right side as well.

    Rick W
     
  5. daiquiri
    Joined: May 2004
    Posts: 5,371
    Likes: 258, Points: 93, Legacy Rep: 3380
    Location: Italy (Garda Lake) and Croatia (Istria)

    daiquiri Engineering and Design

    All this discussion is about test data and validation. Formsys' vallidation page is saying one thing, you are saying another. But it is ok, I can admit that Hullspeed validation page contains errors. Still, they are official, published in their manual and I have called them to intervene and give the missing answers. Until then, maybe you could start sharing your test data, so that by the time they add what they have to say about this issue we will have a complete picture and perhaps will be able to end this discussion in a positive way, with some conclusions.

    One more thing. What is the reason for your heating up around this discussion? It was very calm and instructive until now. Let's keep it that way, please.
     
  6. Guest625101138

    Guest625101138 Previous Member

    You start with the premise that Michlet is wrong and the test data and input data are correct.

    I have learnt from experience that Michlet is usually right and the test data is wrong or the input data wrong. The most significant error in my test data has been actual displacement versus design. The second significant error is water temperature. The third, but less significant, is lift/sink and trim.

    I am not an academic and not interested in publishing for the sake of it - this is my hobby not my business. I am building and testing boats in a real world environment. I have extensive unpublished data that suits my need for comparison - a sample is attached using a calibrated human engine.

    If you want to go into more detail with actual power measurement then you can work through the published archives here:
    http://www.adventuresofgreg.com/HPB/2008_06_01_archive.html
    Work forward from June 1 2008 till September 15th and you will see a whole heap of factors that are worked through regarding boats of useful scale. I do not apologise for it being in a raw form - I have no need to collate it any better way.

    The end result speaks for itself:
    http://www.adventuresofgreg.com/HPB/2008_09_15_archive.html
    The boat was a product of Michlet/GODZILLA.

    This thread is about Michlet. It is an extremely valuable tool for determining the merit of hull shapes and it concerns me when people who have not used it and do not understand it automatically condemn it as something less valuable than questionable test data and even more questionable barrow-pushing software salesmen publishing misleading data.

    Leo is an academic and like most wants everything qualified to the nth degree. For me if it produces a reliable result it is a valuable tool.

    Simply change your starting premise and we will get on a lot better. If you do happen to find a significant error in what Michlet produces compared with real world I will be interested. If you are basing opinion on a software sales pitch of doubtful merit then please confirm the voracity of the source first.

    My intention is to correct your misunderstanding of Michlet and make certain that others are not mislead by your opening premise. Someone in search of knowledge would have an open mind and make their own comparisons/validation first before relying on what software salesmen publish.

    Rick W
     

    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

    I suspect that they used the NPL offsets and monohull input file provided as an example in Michlet. That is unlikely to be the cause of the problem.

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

    Leo Lazauskas Senior Member

    I disagree.
    1. I don't believe that there are any mistakes on the Formsys page except that they have used RW+RH for the Michlet curve and not just RW as for their results. They are not comparing apples with apples. It's as simple as that IMO.

    2. The experimental points are just taken from the Ship Science Report 71 by Molland, Wellicome and Couser. Patrick Couser (who, I believe, worked with Formsys) kindly sent me his own personal copy of that report. However, there are good reasons to doubt the accuracy of some of the experiments with the NPL hulls, particularly the 6a which has a very small draft.

    I am sure Molland et al well know the experimental difficulties. In fact, they re-tested some hulls because there were differences with Insel's earlier experiments using the same hulls. The comparisons of the tests are shown in the Molland et al report.

    3. I see no reason why the offsets would be incorrect in their files or the Michlet file. The hull definitely has a transom, and not a bluff pointed stern. I do take your point that Delftship can give unusual stern offsets however. But that's not the prolem in this case.

    4. BTW, I am not, nor ever have been an academic as such. I dislike the sausage-factory approach of Universities and never finished an under-graduate degree; I have never given lectures in any course, and I have actually spent very little time on-campus, except for the University Bar when I was young. I thought University was mostly about meeting girls - maybe it is different in these modern times :)

    All the best,
    Leo.
     
  9. patc
    Joined: Jan 2010
    Posts: 20
    Likes: 2, Points: 0, Legacy Rep: 36
    Location: Area65

    patc Junior Member

    hi -- I was asked to reply to this forum. I am patrick couser, I studied at the University of Southampton under Dr AF Molland where I continued the work of Dr Mustafa Insel on the calm water resistance of multihulls -- this was back in the late 80's - early 90's last century -- we were looking at hullforms based on the Wigley and NPL round-bilge hullforms -- you can read all about it in Mustafa's and my own thesis as well as several papers published in Trans. RINA and ISP as well as several Ship Science reports that you can get hold of via the University of Southampton.
    just found them on line:
    http://eprints.soton.ac.uk/46442/01/071.pdf
    http://eprints.soton.ac.uk/46441/01/072.pdf
    http://eprints.soton.ac.uk/46407/01/124ShipScience_Report.pdf
    http://eprints.soton.ac.uk/46409/01/127ShipScience_Report.pdf

    I myself tested about 10 NPL hullforms in various cat configs as well as monohull. Mustafa tested some others and the Wigley hulls. Another couple of guys tested a few more variants after me (Lee and D.Taunton, if memory serves). As well as calm-water resistance tests, where we measured the wave pattern with remote wave probes and tried to use this data to calculate the wave pattern resistance, I also did a number of tests in head seas.

    I also implemented the slender body method in Hullspeed -- which actually turned out to follow the research reports of Leo and E.Tuck, rather than the work I did at Southampton. I did however keep the "virtual appendage" method for closing the stransom. I would agree that there is little?any physical justification, other than it appears to give the best correlation to measured wave pattern resistance. There is still a lot I?we dont understand about catamaran resistance, particularly an apprent "viscous" interaction even with widely spaced hulls (I this could well be something else, but it appears to scale as Re and not Fn so Im calling it viscous). In the original work at southampton we did play about with "tweaking" the virtual appendage length to get better matches, however I realise that this may have "tuned" model to NPL-type forms so in Hullspeed we just have a fixed length relative to the transom beam at that waterline. Dr L. Doctors uses a similar method to improve predictions when using a Michel type approach to wave resistance.

    I must admit that the comparisons with Michlet were done way back in ancient history and I dont recall the exact version etc or procedure that I used -- Attached are the files I used.


    It should not be forgotten that, at the end of the day, this approach is a very simple, linear, potential flow model. But it does give pretty good results (at least for comparitive purposes) for a reasonably wide range of hull forms. Even with state of the art viscous resistance solvers, these still need to be "tuned" with tank data for each individual hullform. I see Hullspeed/Michelet as a realtively idiot-proof method, where non-experts can get decent comparitive results without the chance of being really way off target -- a bit similar to strip theory for seakeeping predictions.

    Hope this helps,
    cheers,
    pat

    PS if I can figure out how to attach a file Ill add the raw comparison data I used for Hullspeed etc
     
  10. patc
    Joined: Jan 2010
    Posts: 20
    Likes: 2, Points: 0, Legacy Rep: 36
    Location: Area65

    patc Junior Member

    attachemnt -- I hope

    apparently there should be an attached file?
     

    Attached Files:

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

    Leo Lazauskas Senior Member

    Hi Patrick!
    Thanks for clarifying some aspects for us. We needed another referee here because some players are close to off-side :)

    I do take (a minor pedantic) issue with the use of the term "slender body method".

    The NPL hulls are definitely slender insofar that they have small draft and small beam, but it is confusing to refer to the theory as slender body theory.

    In slender body theory, there is no dependence with depth as there is in Michell's thin-ship theory. The slender body theory of wave resistance (as developed independently by Maruo, Tuck and one other guy I can't remember) was rather disappointing in that it gave negative wave resistance at low Froude numbers for some hulls.

    All the best!
    Leo.
     
    Last edited: Jan 12, 2010
  12. patc
    Joined: Jan 2010
    Posts: 20
    Likes: 2, Points: 0, Legacy Rep: 36
    Location: Area65

    patc Junior Member

    hi Leo -- Good to hear from you, happy new year hope all is well!
    doesnt slender = thin (just more polite :)
    I must admit that I rarely have much spare time to keep up with developments in this area so its good to see that such "simple" methods are still useful when so much effort is going in to non-linear viscous resistance codes.
    cheers,
    pat
     
  13. patc
    Joined: Jan 2010
    Posts: 20
    Likes: 2, Points: 0, Legacy Rep: 36
    Location: Area65

    patc Junior Member

    Sail like an Egyptian

    pushing the limits -- as a matter of interest I tried using hullspeed to estimate the resistance of an encient egyptian sailing vessel I was helping to design:
    http://sunnypowers.perso.neuf.fr/sunnypowershome/na_projects.html#egypt
    We only had extremely rough trial data to go by, but I was pretty amazed by the correlation I got (perhaps it was because there was so much scatter in the trials data that any line would have fit quite well :)
    This paper was presented at the RINA Historic ships conf last November.
    cheers,
    pat
     

    Attached Files:

  14. Guest625101138

    Guest625101138 Previous Member

    Pat
    I have checked the files you posted. There is no apparent error.

    I have attached an unnormalised comparison of Michlet using your in.mlt file with the .csv for the hull against the NPL4a experimental data.

    To me the comparison looks fair - just kidding Leo. If it was any better it would be considered contrived.

    I am at loss to explain how the Formsys Michlet comparison was arrived at. It seems flawed.

    Rick W
     

    Attached Files:


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

    Leo Lazauskas Senior Member

    Happy New Year to you too!

    All's well here in Australia - same as always - appalling heat, droughts, floods, and raging bushfires. No wonder people risk their lives in leaky boats to get here!

    Slender hulls have small beam and small draft compared to length, i.e. like a pencil.
    Thin hulls have small beam compared to length. There is no restriction on draft (except it should not be too small) and Michell himself considered vertical struts with infinite draft.

    In thin-ship theory, sources are distributed over the centreplane of the hull.
    In slender-body theory, the input is the section area curve. There is no distinction between the actual shape of each cross sections. i.e. a line of sources of varying strength rather than a sheet in thin-ship theory.

    Just as a ship should be designed with its mission in mind, so computational methods should be used with what is being modelled in mind. Michell's theory is completely inappropriate for stubby ships, close to the ship where non-linear effects are evident etc etc. Non-linear methods and CFD are complete overkill for thin hulls when we are only interested in the wave resistance, squat or far-field wave patterns. And it is not apparent to me that CFD is actually any more accurate than linear methods for resistance of thin ships.

    The attached pdf compares predictions of Michell's theory + ITTC skin-friction for a Wigley hull. Hullspeed would produce almost identical results.

    Also shown are experiments in deep and finite depth and the predictions of a modern CFD code. This CFD code is an interesting case in itself because it was developed by Stern et al who are very concerned with the proper validation of CFD codes. F. Stern has been instrumental in the establishment of methods to validate CFD and also in a world-wide program to measure resistance of two models at about 30+ towing tanks.

    I have an obvious bias, but I can't see that CFD is doing much better than the simple method for this Wigley hull.

    CFD codes shouldn't have any great problems with meshing the hull - as you know, the hull surface is simple, smooth, and mathematically defined. But maybe I am being naive about that -I actually know little about meshing in CFD.

    As you also know, there is another great advantage that linear codes have over non-linear codes and CFD when it comes to predicting far-field waves. Those methods have to calculate the waves from the ship all the way back to the points of interest in the wave field. Linear codes can calculate the wave elevation at a point anywhere in the field without calculating at intermediate points. Of course, several "hybrid" CFD codes use the linear approximation in the far-field but they then have to perform some numerical trickery to match CFD solutions and linear approximations at boundaries. That's pretty tricky work for non-specialists!

    As you wisely noted, linear methods are more idiot-proof which is a bit of an advantage in itself!

    Leo (occasional idiot).
     

    Attached Files:

    • fd1.pdf
      fd1.pdf
      File size:
      12.6 KB
      Views:
      516
    Last edited: Jan 12, 2010
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.