Sysser 62 hull geometry ?

Discussion in 'Hydrodynamics and Aerodynamics' started by quequen, Sep 9, 2015.

  1. quequen
    Joined: Jul 2009
    Posts: 370
    Likes: 15, Points: 28, Legacy Rep: 199
    Location: argentina

    quequen Senior Member

    Delft hulls and data analysis have been released for public access, but I still can't download the files. Can anyone throw light on the procedure to have access to the files?. Alternatively, does anyone have the Sysser 62 hull and specifications in any 3d format (iges, fbm, dxf, obj, stl, step, etc)?
    The attached file (Ortín Montesinos) is a deeply CFD analysis (made with CFX) of this particular hull at many different conditions.

    https://www.academia.edu/14339768/Hydrodynamic_Modelling_of_Hulls_Using_RANSE_Codes

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

    Leo Lazauskas Senior Member

    Interesting paper.
    I wonder what results he would have got if experimental results were not
    available before he did the predictions. :)

    Uli Remmlinger did some work with this series. I'm not sure if the hull
    geometry is available at his site.

    http://www.boatdesign.net/forums/hydrodynamics-aerodynamics/delft-hull-series-51213.html

    http://www.boatdesign.net/forums/design-software/new-lines-prediction-program-45601.html

    I have some files in Michlet format he sent me last year. I'm not sure what
    you want, but he might be able to provide you with something if you send
    him a message.
     
  3. quequen
    Joined: Jul 2009
    Posts: 370
    Likes: 15, Points: 28, Legacy Rep: 199
    Location: argentina

    quequen Senior Member

    Thanks Leo, I emailed delft many times, no response yet... I'll ask Uli as you suggested.

    Here one of the videos related to Montesinos paper:


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

    Remmlinger engineer

    Don't have much time at the moment.
    Uli
     

    Attached Files:

  5. quequen
    Joined: Jul 2009
    Posts: 370
    Likes: 15, Points: 28, Legacy Rep: 199
    Location: argentina

    quequen Senior Member

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

    Remmlinger engineer

    Thanks for the link!
    I agree with Leo, we do not know how much the results were tweaked to match the experiments. There is no information in the paper how the influence of the 3 sand strips was modeled. Was the boundary layer forced to be turbulent right from the beginning? How was the turbulence length scale chosen? You can change the results a lot with these parameters. It would have been nice if the author had shown the influence on the results of a variation of these parameters.
    Nevertheless, the results are quite good, but the price is high. 40 hours on 12 computer-cores for each data point! And the work seems to be so demanding, that it earns you a master in engineering!
    As long as the hull is similar to the DSYHS, I prefer my regression, see attached diagram. The regression method takes only milliseconds on a one-core computer.
    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

    There are many similar CFD papers to that MSC thesis appearing in journals.

    It's extremely unlikely that anyone would be able to replicate the results
    because they depend on the mesh used to describe the hull for the particular
    computer code. However, I like some CFD papers because they (necessarily!)
    contain experimental data. Without that data many are worthless.

    As you said about the above paper, there is no mention of how the sand
    strips were modelled. That is a very serious flaw and I'm a bit surprised a
    supervisor did not get the student to include some statement about it.
     
  8. Jose Luis Ortin
    Joined: Jan 2014
    Posts: 10
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: Southampton, UK

    Jose Luis Ortin Junior Member

    Hello All,

    First of all thanks for reading the paper and for your comments, it is always good to have some feedback.

    As someone mention before this is an academic paper, I tried to generate a methodology to reproduce towing tank results as close as possible from almost a blank paper for this specific software. In that regard the main point was to try different meshing and setting up configurations to understand how the different changes affected the results. Having the results allowed precisely to develop this methodology,without them this would have not been possible. The point was not to reproduce the results but to see how far we could get in developing a methodology. That is why it makes no sense to tweak the results, we just tried different setting ups until we got something sensible; this was a serious research project which included a lot of previous preparation and reading. When you submit the project the people who evaluate it have access to the results files and trust me that you do not want to tweak them. All the parameters were carefully set up by reading previous work of other authors (please have a look at the references).

    We tried as many different configurations as we could but we did not have the time to test them all. Nevertheless your comment regarding the sand strips are right, I actually thought I had mentioned it in the report. We tried a transitional model in order to model the laminar part of the flow and it failed due to the fact that yplus values of 1 are needed to correctly apply this kind of transition. This is one of the reasons why the resistance is over predicted; all the flow along the hull is treated as a turbulent flow which in a model scale make a considerably difference in the results, even with the sand strips in the experiental model. I wanted to make a deeper research of these transitional models but we did not have time for more, the idea was to pass the baton to the next student.
    Regarding the turbulence intensity I do mention in the report that a 5 % turbulence intensity was chosen (page 22) along with the SST turbulence model. We could not find any turbulence data from Delft, that is why we chose that default turbulence intensity (as recommended by Ansys CFX), leaving the other turbulence parameters to be automatically computed. This was more prudent than choosing just a random length scale and intensity. Please keep in mind that in this kind of academic paper you have to justify everything, either in the report or in the subsequent presentation.

    I am keen to help in all I can, that is another of the main purposes of doing an academic research. Please email me if you need any more information (j.l.ortin@hotmail.com).

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

    Remmlinger engineer

    Thank you Mr. Ortin for joining this discussion!
    It does not happen very often, that an author is available for questions!
    First of all I want to stress that it was not my intention to run your thesis down.
    The discussion became so critical, because many vendors of commercial CFD-programs claim, that CFD can replace towing tank experiments. The client is told, that he can design a competitive sailing yacht just by using CFD in a straightforward way. Reading your thesis shows that this claim is not true. Instead a lot of fundamental mathematical work is necessary.
    In your post you are making the point that your main intention was to develop a methodology for the resistance prediction of sailing yachts. It would be interesting to know, if this methodology could be cast into some "rules", that would help to achieve reliable CFD-results even in the case, when no towing tank tests are available.
    I personally do not believe that CFD can replace the towing tank. Instead I see the big advantage that CFD can compliment the tank tests by providing an artificial flow visualization. Therefore it is a pity that you did not include streamline pictures in your thesis. I would be especially interested in the angle of attack of the rudder under heel and leeway (is it negative?).
    Thanks for sharing
    Uli
     
  10. Leo Lazauskas
    Joined: Jan 2002
    Posts: 2,696
    Likes: 155, Points: 63, Legacy Rep: 2229
    Location: Adelaide, South Australia

    Leo Lazauskas Senior Member

    Welcome, Jose!

    I appreciate that you were investigating methodologies for your
    thesis.

    Maybe we have different ideas of what "tweaking" means.

    You say that you tried "to reproduce the towing tank results as
    close as possible".

    That is what I mean by "tweaking" results. (I certainly don't mean
    that you were scientifically dishonest in any way). For me it is
    that process of adjusting meshes or other factors until there is
    good agreement with known measured values that makes many CFD papers
    almost worthless from a scientific aspect. Those papers are often
    little more than marketing exercises because the code is being used
    as an elaborate fitting mechanism. The underlying physics of the
    CFD code is not very important when it is used in that way; the same
    good correlations could be achieved with a fancy statistics package.

    Good luck with your future studies!
     
  11. Jose Luis Ortin
    Joined: Jan 2014
    Posts: 10
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: Southampton, UK

    Jose Luis Ortin Junior Member

    Hello Eli and thanks again for your comments, your feedback is really helpful!.

    I believe that towing tanks are necessary but just for very specific projects and for CFD validation purposes. It is a fact that nowadays towing tank testing is by far less common than 10 years ago. I agree with you though, the numerical prediction of hull resistance it's something rather complicated that takes some time and a considerable amount of expertise is needed to obtain meaningful results, it is not a straightforward tool to optimise hull shapes. The main point of the validation project was precisely to get those "rules" that could be later applied to other problem when experimental results are not available.
    I wanted to include more pictures but the graphs and numerical results were more important for the supervisor.
    The angle of attack of the rudder coincides with the yaw angle of the hull.

    Cheers!
     
  12. Jose Luis Ortin
    Joined: Jan 2014
    Posts: 10
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: Southampton, UK

    Jose Luis Ortin Junior Member

    Hello Leo, it seems that you don't really understand what a CFD validation project means, this is a very basic concept within CFD. Before using any CFD code for any problem the mesh and the methodology must be validated, otherwise the results are worthless. In this case I validated a methodology to calculate the hydrodynamic resistance of hulls; applying this methodology with this particular code I can be quite confident in predicting the resistance of any other hull shape.
    The main difference between a fancy statistical method and a CFD code is that once the latter is validated it can be applied to any other hull shape, which is hardly achievable by any systematic series or regression methods.
    This is why you find many similar papers. They are not mere marketing exercises, it is just a necessary step before attempting any real problem where the solution is unknown. They do have a scientific value, these are normally evaluated by experts in the field and presented in recognised hydrodynamic conferences. You should be more respectful not just with the work made by the authors but with the fact that they are sharing it publicly.

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

    Leo Lazauskas Senior Member

    I have great respect for Fred Stern and his group who were instrumental in
    setting up formal V&V procedures for CFD after a lot of criticisms were
    levelled at the field about 10-15 years ago. The EU were also very critical of
    some CFD practitioners and practices and demanded that certain standards
    be adhered to.

    My objection is to papers that appear with results that can't be reproduced.
    They are of the type that say little more than we got these results with
    Fluent/Starccm/etc with no explanation of the mesh used and how results
    varied with coarser meshes or a different grid. That type of paper deserves
    as much respect as junk mail.

    CFD is a splendid tool when used properly and appropriately. But there are
    also charlatans in the industry who are as trustworthy as used car salesmen.
    That's why Fred Stern and the ITTC instituted formal V&V procedures, so read
    up on the history before demanding that I be respectful of crooks and
    conmen.
     
  14. Jose Luis Ortin
    Joined: Jan 2014
    Posts: 10
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: Southampton, UK

    Jose Luis Ortin Junior Member

    Fair enough, it's good to clarify an opinion.
    Now I coincide with you. It seemed by your previous comments that you were judging everybody who has something to do with CFD. I guess that those who don't explain too much their methodologies are more interested in a personal promotion than in any academic research.

    By the way thanks! this quote made mi day ;)

     

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

    Remmlinger engineer

    This is hard to believe. All towing tests at heel with zero leeway show, that there is a rudder-sideforce even at zero rudder angle. The reason is the distorted non-symmetric flowfield, caused by the hull. With additional leeway there is the downwash from the keel which additionally changes the angle of attack. The angle of attack will most likely not be Beta*COS(Phi).
    May be by "optimizing" your resistance results you have introduced several errors that cancel each other?
    This is one of the reasons why I only trust tank experiments.
    Uli
     
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.