Table of Offsets import bugs?

DELFTship forum Hull modeling Table of Offsets import bugs?

Viewing 3 reply threads
  • Author
    • #44268

      When I export a Table of Offsets from a simple, well-behaved model of something like a small sailing dinghy, I get a table of offsets that I believe I understand and that looks like it appropriately represents the model.

      However, if I then import that same file into a new project, the result seems to produce some errors:

      1. The aft surfaces seem distorted due to some station-waterline offsets close to keel centerline being confounded with some aft contour (waterline-centerline intercept) points.
      2. The aft contours don’t close along transom to the deckline.
      3. The forward contours don’t close the forefoot and stem.
      4. Otherwise, the surfaces between waterlines and stations appear okay.

      I’ve also tried various combinations of the “read flags” in the header of the offsets file with these results:

      1. “Read flat of bottom” is set to “0” (off) as the export file default.  As expected, when the file is imported into a new project with this setting, the model does leaves out the points at the bottom of the stations near mid-ship. Therefore, no faces are generated between the keel and the lowest waterline.  Not expected, however, setting the read flag to “1” returns an error that prevents the file from importing.
      2. “Read aft contours” is set to “1”  in the default export.   When the file is imported, the problems described above result.  Changing the flag “0” creates other strange results.  Points in the “region of zeros”, below the aft keel, seem get confounded with points on the hull surface to create various weird surfaces aft and below that don’t represent the model.

      Am I wrong to conclude these are bugs with import parsing?

      Does anyone know of workarounds for these issues?

      I’ve enjoyed stumbling up the DELFTship learning curve over several years of on and off use to generate and learn from several dozens of models I’ve created.  Getting over the problems outlined here will help me to significantly streamline my workflow in a small study project I’ve entered into recently.

      Thank you for your help.

    • #44279

      Most of the time the errors you are referring to are either configuration errors of the table of offsets (ToO), or simply related to the fact that a ToO tries to capture complex 3D geometry in a strictly rectangular grid of intersection points. Behavior of a surface in between those intersection points is unknown when importing a ToO, while at the same time the simple structure of a ToO makes it impossible to include that crucial information.

      It’s difficult to tell whether these are bugs or not without having access to your table of offsets file
      Are you willing to share your file with us?

      • This reply was modified 2 weeks, 3 days ago by Marven.
      • This reply was modified 2 weeks, 3 days ago by Maarten.
    • #44282

      Thank you for the response.  I’ve attached a representative file.

      The file is a table of offsets export from a model.  I checked that it behaves as I described in the first post when imported into a blank new project file.

      The model shape seems straightforward and the export function seems to capture all the expected values for station and waterline grid points, deck (shear) line, and contours (fore, aft, and bottom).  The import function does not seem to use all the information the export function delivers.

      Thank you for taking a look at these additional clues.

    • #44287
      Denis Pratt

      It might have something to do with your intersections.

      if there is no waterplane at zero draft you will get an  error, likewise for the further points fwd and aft for your stations.

      Not needed to use the offset table feature for my work, but tried the technique on a hull I designed with no station at the extreme lengths and without a waterline at the bottom of the hull and found those problems.


Viewing 3 reply threads
  • You must be logged in to reply to this topic.