how to design in 3D

Discussion in 'Software' started by 1gerry, Jul 27, 2013.

  1. 1gerry
    Joined: Jul 2013
    Posts: 15
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: Canada

    1gerry Junior Member

    Hello,

    I am starting a new hull in rhino.

    Altough I have lots of 3D experience, when I loft the surfaces using sections I get some odd results (ie, lines plan on top view makes a curve that bends inward, instead of a nice almost straight line)

    How do you start your drawings? Do you draw the sections first and then connect with lofted surfaces, or do you draw the lines in each plane and then project to 3D to create the surfaces?

    See the attached image for an example effect of what I get (in red is my simulated result: the curve that bends inwards. I drew this one by hand over a random plan, just to show the exagerated effect. This bend is replicated in all other lines)

    Also, is this bad (having this kind of curve instead of a straighter one) ? What would be the effects of such a hull??
     

    Attached Files:

  2. DCockey
    Joined: Oct 2009
    Posts: 4,243
    Likes: 184, Points: 63, Legacy Rep: 1485
    Location: Midcoast Maine

    DCockey Senior Member

    Please provide the side and end views in addition to the top view.

    Are the curves which appear in the top views the isolines of the surface, or curves you created?

    Are you trying to design a new shape or are you starting with an existing set of lines?
     
  3. DCockey
    Joined: Oct 2009
    Posts: 4,243
    Likes: 184, Points: 63, Legacy Rep: 1485
    Location: Midcoast Maine

    DCockey Senior Member

    How much Rhino experience do you have?

    Do you have other CAD experience? If so what software and have you done freeform design or only "mechanical" type design or architecture design?

    Have you drawn a set of boat plans with pencil and paper?
     
  4. DCockey
    Joined: Oct 2009
    Posts: 4,243
    Likes: 184, Points: 63, Legacy Rep: 1485
    Location: Midcoast Maine

    DCockey Senior Member

    Typically for a new simple, round bottom shape I start by creating a mid-section, transom, keel, stem and sheer. I use the same number of control points of the mid-section, transom and stem, and for the keel and sheer. Then I generate an initial surface using the NetworkSrf or Sweep2 commands. Initial sections, waterlines and buttocks are obtained using Contour. I look at the lines and surface, and also inspect the surface including using Zebra and EMap (fluorescent tube).

    Next I'll adjust the previously generated curves as desired, and usually add a section between the stem and mid-section, and one between the mid-section and transom. Rather than creating new curves I'll copy the mid-section curve to the new locations along the keel, then use Scale1D to adjust it so fits to the sheer. Then I'll modify the shape of these new sections as desired by moving the control points for the curves. Then generate another surface using NetworkSrf or Sweep2. Revised sections, waterlines and buttocks are obtained using Contour. I look at the lines and surface, and also inspect the surface including using Zebra and EMap (fluorescent tube). I also run the basic hydrostatic calculations using Hydrostatics.

    Then I iterate adjusting the curves, generating surfaces and lines, and inspecting the resulting surface and lines and doing the hydrostatic calculations. I may add some additional sections or longitudinal curves. I keep iterating, adjusting the curves, until the surface is very close to what I want and the hydrostatics are satisfactory.

    Only after I have a "very close" surface do I start adjusting the surface control points to fine tune the surface.
     
  5. 1gerry
    Joined: Jul 2013
    Posts: 15
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: Canada

    1gerry Junior Member

    Hi David,

    Thanks for all the information.

    I see there is a lot of tweaking.

    I am an industrial designer, so I am quite familiar with complex surfaces.

    Do you use curvature combs to analyse the resulting surfaces, and see if there are kinks anywhere?

    Attached, a few images to illustrate the model as you asked. This is just a test model as I am for now, just exploring what happens with the curves.The bending inwards I described above seems to often be present so I'd like to know what it does, and why should I stay away (or not) from curves like these.

    Thanks again, and looking forward to your comments.
     

    Attached Files:

  6. DCockey
    Joined: Oct 2009
    Posts: 4,243
    Likes: 184, Points: 63, Legacy Rep: 1485
    Location: Midcoast Maine

    DCockey Senior Member

    Yes, I use curvature combs but use zebra stripes and particularly EMap with the single fluorescent tube more, particularly for fine tuning the surface. The EMap reflection is very sensitive to kinks provided the render/display mesh is fine enough. Go to properties and adjust it to Smooth and Slower at a minimum.

    The shape of the waterline curves aft is the result of the shape of the bottom with the deadrise decreasing aft and the keel sweeping up. Experiment with different shapes and it will become clearer.
     
  7. Mike Graham
    Joined: Feb 2013
    Posts: 67
    Likes: 6, Points: 0, Legacy Rep: 54
    Location: Maryland

    Mike Graham Junior Member

    It's usually best to create hull models using surfaces with relatively few control points and to edit them directly. It's a bit trickier to get used to, but it makes it a lot easier to create a fair model with few hiccups.

    This video is an example of someone using direct surface modeling to create a hull.

    Lofting is best saved for when you already have a lines plan and want to reproduce the hull as best you can, and even then it can be tricky to get good results.

    It's difficult to say what one should compare it to--each change has effects on multiple things--but in general I would not expect that kind of concavity to be beneficial from a performance perspective or a construction perspective, but not expect major problems. It may inflate wetted area, exacerbate local buckling of the structure, and be hard to fabricate. If it's really slight, it should not have a huge effect at all on performance.

    Engineering is done with numbers, though, and time after time we've seen that educated guesses on things like hydrodynamic performance can be very wrong. I have seen high-fidelity simulation hull optimization result in concave areas in the shell, but in a much more intuitive explicable area.
     
    Last edited: Jul 28, 2013
  8. DCockey
    Joined: Oct 2009
    Posts: 4,243
    Likes: 184, Points: 63, Legacy Rep: 1485
    Location: Midcoast Maine

    DCockey Senior Member

    The video link does not appear to work. What's the title of the video on YouTube?

    I think he as referring to creating a surface using the Loft command in Rhino which creates a surface by interpolating between curves, not "lofting" as in drawing a set of lines full scale in preparation for building.
     
  9. DCockey
    Joined: Oct 2009
    Posts: 4,243
    Likes: 184, Points: 63, Legacy Rep: 1485
    Location: Midcoast Maine

    DCockey Senior Member

    For a V-bottom hull like 1Gerry shows I use a different approach for creating the hull shape than I outlined above for a round bottom hull.

    First I create initial versions of the sheer, chine and keel curves, with the keel curve running up to the chine.

    The simplest way to create initial surfaces for a V-bottom hull is to use the Rhino Loft command between the keel and chine, and between the chine and sheer. If developable surfaces are desired select "developable" for the style in the pop-up box of the Loft command. Remember that straight sections usually do not result in developable surfaces. The shape of the sheer, chine and/or keel curves may need to be adjusted so that developable surfaces are created for the entire length of the curves. I use the usual techniques for evaluating the surface; Emap with single fluorescent tube, Zebra stripes, curvature combs, curvature shading. I also create lines, particularly sections using the Rhino Contour command. Hydrostatics are calculated and checked. Then the shape is revised by modifying the sheer, chine and keel curves as desired and generating a revised shape. I iterate until satisfied.

    The disadvantage of the simple method outlined above using the Loft command is that Loft will generally form conical portions of surface at the ends of the curves. The curvature of these conical surfaces may become large near the apex of the conical sections which in theory can cause problems. In practice when building with plywood it may be okay. Because of this I now use a somewhat different method for V-bottom hulls with developable surfaces which I'll describe in another post.
     
  10. Mike Graham
    Joined: Feb 2013
    Posts: 67
    Likes: 6, Points: 0, Legacy Rep: 54
    Location: Maryland

    Mike Graham Junior Member

    Oops. I edited the post to fix the link

    Rhino3d Demo Ship Hull Surface Control Point Modeling - http://www.youtube.com/watch?v=hgjxO5yuRxg


    Yes, I was referring to using the Rhino Loft command. Depending on the options you use, this can tend to give a rather ugly or impractical surface. The best lofted surfaces tend to make use of the "loose" option, which is at odds with your sections/butts actually being adhered to in the process.

    For hulls designed from scratch with any sort of interesting shape, using a surface which relies on a small number of control points tends to give the nicest, fairest results in my experience and in what I've been told.
     
  11. Mike Graham
    Joined: Feb 2013
    Posts: 67
    Likes: 6, Points: 0, Legacy Rep: 54
    Location: Maryland

    Mike Graham Junior Member

    To quote the Orca3D manual on the subject

     
  12. DCockey
    Joined: Oct 2009
    Posts: 4,243
    Likes: 184, Points: 63, Legacy Rep: 1485
    Location: Midcoast Maine

    DCockey Senior Member

    Thanks for updating the link. I've watched that one before, and I've seen also seen that method demonstrated by Cliff Estes. For the type of hull shape shown with a parallel mid-body it appears to work well, particularly if the exact shape of the hull is not critical. The trick is to know where to position the control points.

    I agree with using the minimum number of control points consistent with the desired shape. There can be a tradeoff though between minimizing control points and obtaining the desired shape. I sometimes wonder how many of recent sailboat and powerboat shapes are the result of minimizing control points and just accepting the result.

    For the methods I use the resulting surfaces are simpler if all the transverse "section" curves have the same structure (number of control points, etc), and the longitudinal curves have the same structure.

    FitSrf, FitCrv and Rebuild are very useful for simplifying surfaces. Occasionally I'll add a row or column of control points to a surface to increase control in an area, particularly where tight curvature may be desired.

    I don't use the Loft command for "lofting" a set of surface station curves. Instead I'll add longitudinal curves, make sure they are the desired shape, and then use Sweep2 or NetworkSrf to create surface. Recently I've been using Patch command to obtain a surface from curves but there are some techniques needed beyond simple application of Patch to obtain a good surface. Perhaps I'll describe that method sometime. The Loft command can be useful though for hard chine hulls as I described above.
     
  13. DCockey
    Joined: Oct 2009
    Posts: 4,243
    Likes: 184, Points: 63, Legacy Rep: 1485
    Location: Midcoast Maine

    DCockey Senior Member

    I stongly disagree with the first sentence of the quote.

    There are many, not only two, basic approaches to modeling hulls in Rhino.

    The two methods described are commonly used ones, and are probably the ones most commonly presented. But they are not the only ones. The methods I use do not fit either of those described.
     
  14. 1gerry
    Joined: Jul 2013
    Posts: 15
    Likes: 0, Points: 0, Legacy Rep: 10
    Location: Canada

    1gerry Junior Member

    Thanks all for your comments!

    Do you usually tweak points in the longitudinal lines, or more in the vertical (section) lines?

    I am finding it hard to move points arround, as I can't see the result until after the function, and then it's a lot of back and forward...

    Thanks again.
    I'll post something soon
     

  15. DCockey
    Joined: Oct 2009
    Posts: 4,243
    Likes: 184, Points: 63, Legacy Rep: 1485
    Location: Midcoast Maine

    DCockey Senior Member

    Depends on what I want to change. Usually I spend some time and get a sheer curve which I like and then don't change it until the final tweaking. Then it depends on where the revisions are needed. I usually have 3, maybe 4 stations between stem and stern. I'll concentrate on them. If I've added longitudinals between the keel and sheer I'll revise those so the pass through the stations (Exact isn't absolutely necessary by they should be close.)

    For a shape with a chine such as you've shown I have separate surfaces for the topsides (between sheer and chine) and the bottom (between chine and keel).

    Nudge the points with the arrow keys! Read the Nudge section in Help and adjust the size of the nudges in Properties under Modeling Aids.
     
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.