NetworkSrf v. Sweep2

Maybe someone can explain this general problem that I have with Sweep2 when using table of offset data. That is, accordion pleats. I have read many posts and seen several Youtube video decrying the use of NetworkSrf.

Here I have table of offset data:


The curve at the top and bottom of the point set have been faired to get rid of rounding errors and then I regenerated the Y coordinates by intersecting. Then I have drawn the vertical curves through the new points in the original points. I have spaced the vertical curves every four frames, except the end where I did 3 frames.

Here I have done networkSrf:


The highlighted curves are intersections I did at the waterlines matching the points. Note that these curves run close to the original points that I did not include in the curve set.

Here have used the same curves with Sweep2.


Now the intersections waterlines are wavy.

This is not always the case. Here have three surfaces joined. The forward surface is a tangency zone, so I used edgesrf. The surface above the knuckle is a sweep2. It’s sections are mostly lines.

I’m, still trying to figure out what I am going to do for the dreaded corner (which in this case is actually 4-sided). Sweep2 gives the accordion pleats there.

  1. What causes sweep2 to create accordion pleats?
  2. Does this illustrate use the right too for the right job?

Hello - it is hard to say without having the input curves.

-Pascal

Here is an example of the same:
Problem Sweep.3dm (3.6 MB)

I have the same frames using Sweep2 and NetworkSrf. A cross section shows the former has a wavy navy surface.

Hello - with this many cross sections, Loft seems like a much better bet - have you tried that?
Your file is in inches, and the tolerance is .001 - this seems excessively tight to me, and I agree thatSweep2 seems not to do a great job here.

-Pascal

The low tolerance is from the template I copied the curves into.

I have wondered if I am missing some trick. I typically have been using RebuildCrvNonuniform on groups of frames to get their control points aligned but that does not seem to work.

I am trying to apply all the things I have learned over the years to this. The only place where I have gotten sweep2 to work is alone the surface at the upper left where the sections are lines for the most part then transition to simple curves aft. Probably not coincidentally, that is the type of surface that appears in don’t-use-networksrf video examples of boat hull.

The reality is the shapes on a ship tend to be much more complex. In this unfinished example there are knuckles (horizontal lines at the left in the lower image), there are transitions from three inflection points to none, the are sections with straight surface. When the surface has a lot of bends with transitions, Sweep2 does not appear to work well.

Hello - for the curves in your file, Loft gives the best results from what I can tell.

-Pascal

I generally get good results with loft; except that the edge tend not to match what I need to the adjoining surface.