Hello everyone
I am a newbie to Rhino, may I ask how to restore the two graphics in the photo in Rhino? For example, if I draw a plane or a vertex on the top, the sweep line of the graph, and the wave curve at the bottom, what operations do I need to perform to create such a untrimmed surface? Because I tried split graphics, I got a trimmed surface, I need an untrimmed surface for input in the grasshopper operation. Thanks in advance
Tianci
2 rail sweep, curve a the top, curve at the bottom, and a profile curve to control the surface shape.
network surface 
I try to avoid network surface if possible as it makes really really complex surfaces where a simpler solution can be had with other methods.
I look at most of the solutions as boiling down to basically the same thing. But network srf seems to be where everything leads – imo.
If what you mean by ‘complex’ is the control point density, then that can be simplified afterward of course.
I find that netwrksrf’s can actually simplify the nurbs composition discovery process in seemingly difficult scenarios.
I guess it all depends on the user’s strategies.
Doesn’t actually work in this case, while in theory it seems easy enough, in testing with the OP’s original input rails and profile, Sweep2 is just not capable of doing this without making a self-intersecting mess. NetworkSrf however does seem to make a surface, will put that in another post.
not the first time I’ve been wrong… 
Here is an experiment I did - in the file below, which used the original curves above, I just moved them to be symmetric around 0 for easier manipulation - I made a revolved trimmed surface and an untrimmed NetworkSrf. I was not able to get decent results from either Sweep2 or Loft, maybe someone is better than me can do it…
The revolved trimmed surface looks very nice…
The NetworkSrf looks sorta OK at first glance…
…until you look closer and turn Zebra or look at the V direction curvature graph…
UntrimmedExample.zip (11.2 MB)
Edit:
I actually did get a bit better Loft surface by using some intermediate profiles and rebuilding them all the same. But the lower edge does not match the rail exactly and the Zebra is not anywhere near as good as the revolve. Might be useful in some cases though.
I worked some magic on it.
The density is a bit inevitable for it to be untrimmed. The secret is how to add it.
UntrimmedExample_emod.3dm (6.5 MB)
I had to use ‘savesmall’
I kind of rushed it, so I’d probably have to mess with it longer to address a few remaining ripples.
I did use your:
as a pulling object for a curve network at one point.
So, I’d have to play around with the network I pull to it, and the condition or density of that object too. And maybe some other tricks.
The remaining ripples might be where the network is still too dense, and smashes together real tight.
I can see how the U direction squiggles a bit, so I’d have to try a trick where I extract only the V direction and get Rhino to recreate the U direction. And probably the same for the V and repull the network again.
I’m not sure the ripple can be removed 100% though. that’s kind of a crazy surface to make untrimmed lol.
The trick of going back n forth with most U or V network while utilizing “original boundary”, is a way to get each direction to ‘relax’, and the finally using a smoother pull surface template to repull it – ultimately getting a final untrimmed version with minimal ripple…
I would use RailRevolve command.
…there’s two profiles and two rails 
theoretically. Unless you trim it.
I guess it all depends on ‘design intent’ and/or geometry conformation dimensions and tolerances.
You can import that image as a pictureframe, and literally trace it, and create the theoretical untrimmed surface in many ways, especially if accuracy isn’t an issue.
Doing it in GH is a whole nother matter.
More than two sketches would help determine the actual 3D intent.
