Why is it?

Discussion in 'Software' started by Tad, May 14, 2010.

  1. Tad
    Joined: Mar 2002
    Posts: 2,303
    Likes: 185, Points: 73, Legacy Rep: 2281
    Location: Flattop Islands

    Tad Boat Designer

    Why is it that exporting an iges file (one surface) from Freeship....it arrives in Rhino in six pieces? Is there a way to tell either Freeship or Rhino not to do this and just leave it as one surface, or to only split surfaces at joints (chines, transom, etc.)?

    Thanks.....
     
  2. conceptia
    Joined: Nov 2009
    Posts: 203
    Likes: 7, Points: 18, Legacy Rep: 97
    Location: Houston

    conceptia Naval Architect

    this happened to me also, when I tried to import the iges into Solid works..
     
  3. Guest62110524

    Guest62110524 Previous Member



    that happens rhino to maxsurf too, you had best ask guru andrew mason from Maxsurf or maybe search the rhino help on net, frustrating eh
    i have found that compound surfaces from rhino(squished) measure same as in maxsurf, that was good because one always frets with 4000 dollar plates, I will try it from freeship, anyways some specialist will be along to help you will see
    I did this one in 07 it came from freeship and plate was one, I cant remember how I only have the rhino files now
     
    Last edited: Jun 21, 2010
  4. Andrew Mason
    Joined: Mar 2003
    Posts: 397
    Likes: 17, Points: 18, Legacy Rep: 206
    Location: Perth, Western Australia

    Andrew Mason Senior Member


    As I understand it, and maybe Martijn can clarify this, is that this problem is because Freeship uses subdivision surfaces, which are not the same as the NURBS surfaces used in Maxsurf and Rhino, Fastship, Autoship etc. I assume that the conversion into NURBS surfaces results in the breaking up of the subdivision surfaces into smaller pieces.

    Maxsurf and Rhino should be able to do lossless transfer of surfaces in NURBS format. If you are getting Rhino surfaces fragmenting when read into Maxsurf you may want to check your Rhino export settings.
     
    1 person likes this.
  5. marshmat
    Joined: Apr 2005
    Posts: 4,127
    Likes: 148, Points: 0, Legacy Rep: 2043
    Location: Ontario

    marshmat Senior Member

    IIRC, the IGES export in Freeship initially tries to represent the hull as many small IGES surfaces, which it then combines into a few large surfaces. At some points- mainly where there are irregularities or discontinuities in the control net- it fails to combine the small IGES surfaces.

    There's no setting in Freeship that can fix this; I'm not sure if the algorithm has been improved in Delftship Pro. It can be (mostly) corrected in Rhino:
    - After importing, make a copy of the surfaces in question on a different layer
    - Use the MergeSrf command to connect the damaged edges
    - Use the RemoveControlPoint command, if necessary, to get rid of the extra control points so that the new surface's control net matches that of the copy you made earlier. (If you don't do this, you sometimes end up with a high control point density near the damaged edge, making it hard to do any final fairing.)
     
  6. Tad
    Joined: Mar 2002
    Posts: 2,303
    Likes: 185, Points: 73, Legacy Rep: 2281
    Location: Flattop Islands

    Tad Boat Designer

    Thank you everyone for the input.....matt I'll give your fix a try....

    It seems the fragmentation is usually in areas of higher curvature, the hull will split into two sections along the turn of bilge, and then there are like 6 little surface fragments in the forefoot....

    An example of what the iges surface looks like in rhino....this one appeared as 27 surfaces.......

    futurecrusier38view.jpg
     
  7. marshmat
    Joined: Apr 2005
    Posts: 4,127
    Likes: 148, Points: 0, Legacy Rep: 2043
    Location: Ontario

    marshmat Senior Member

    Tad: Does the Freeship model for that one have a non-rectilinear control grid? (ie, are there points where only three, or perhaps five, edges meet?) That sort of thing can really confuse the IGES exporter, despite working perfectly in subdivision surface modelling....
     
  8. Tad
    Joined: Mar 2002
    Posts: 2,303
    Likes: 185, Points: 73, Legacy Rep: 2281
    Location: Flattop Islands

    Tad Boat Designer

    matt....

    I see there are couple of "non-rectilinear" sections in the net forward, but aft its very simple.....it's just (to me) that there's a reverse curve (wineglass) in the bottom aft....which seems to have broken the iges surface......

    futurecruiser38net.jpg
     
  9. Tad
    Joined: Mar 2002
    Posts: 2,303
    Likes: 185, Points: 73, Legacy Rep: 2281
    Location: Flattop Islands

    Tad Boat Designer

    Err....now I see there's another partial in the topsides aft.....I'll try fixing those and see what happens....
     
  10. Tad
    Joined: Mar 2002
    Posts: 2,303
    Likes: 185, Points: 73, Legacy Rep: 2281
    Location: Flattop Islands

    Tad Boat Designer

    Okay...that helped a bit....down from 27 to 6 surfaces....the highlighted part is all one...then 5 just in the bow??

    futurecruiser38revnet.jpg
     
  11. Martijn_vE
    Joined: Apr 2005
    Posts: 254
    Likes: 24, Points: 0, Legacy Rep: 401
    Location: Netherlands

    Martijn_vE Marine software developer

    Tad,

    The face in the middle of the red circle is a triangular one and is split into 3 NURBS patches.
     

    Attached Files:

  12. marshmat
    Joined: Apr 2005
    Posts: 4,127
    Likes: 148, Points: 0, Legacy Rep: 2043
    Location: Ontario

    marshmat Senior Member

    One might summarize the root of the problem as follows: NURBS modellers expect the control grid to have an essentially rectangular topology, ie. each row must have the same number of points, each column must have the same number of points, each patch between control points must be a quadrilateral (albeit often a very distorted one). Delftship's subdivision model allows missing points, extra points, and triangular or other unusual shaped faces of the point grid.

    If the NURBS requirements for control grid topology are adhered to in the subdivision surface model (ie. purely quadrilateral grid faces, uniform row and column point counts), I find that there tend to be no problems in the conversion.

    Any deviation from the rules in the subdivision model will lead to a face, however small, that the subdivision-to-NURBS conversion algorithm can't build as part of the same NURBS surface as everything else. So everything adjacent to that patch also has to be treated separately, and you get this multiple-surface fragmentation.

    In Tad's model, I believe one extra control point inserted on the forefoot, to turn the triangular face into a quadrilateral one, should fix the problem without any noticeable change to the geometry.

    (Martijn- correct me if any of this doesn't sound quite right.....)
     
  13. Martijn_vE
    Joined: Apr 2005
    Posts: 254
    Likes: 24, Points: 0, Legacy Rep: 401
    Location: Netherlands

    Martijn_vE Marine software developer

    Totally correct.
    I can add that any face with a number of points other than 4 will be split in a number of NURBS patches equal to the number of controlpoints of that face (3 for a triangular face, 5 for a face of 5 points and so on)
     
  14. Tad
    Joined: Mar 2002
    Posts: 2,303
    Likes: 185, Points: 73, Legacy Rep: 2281
    Location: Flattop Islands

    Tad Boat Designer

    Excellent work guys, thank you....

    The iges export is now down to 5 surfaces, two are highlighted in the image below and there are 3 in the small area at the forefoot.

    futurecruiser38revnet2.jpg
     

  15. marshmat
    Joined: Apr 2005
    Posts: 4,127
    Likes: 148, Points: 0, Legacy Rep: 2043
    Location: Ontario

    marshmat Senior Member

    Getting close.... ;)
    It might take a few tries to find the right order in which to do them (so work on a copy), but if you can MergeSrf the little bits in the forefoot into one surface somehow, it'll then MergeSrf nicely with the rest of the bow.... which should then MergeSrf easily with the rest of the hull. A row or two of extra control points might be created as you do this, which would have to be removed before the next merge.

    Or, you can just join edges and leave the thing as a polysurface.... which works, although it's not as elegant.....
     
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.