Sweep2 v. NetworkSrf and Edge Matching

There are folks here who condemn the use of NetworkSrf as a tool for the lazy.

I am finding that the alternative of Sweep2 creates the problem of not tracking the rails closely.

Here I have an area I need to fill. The upper edge is defined by a surface. The lower edge is an arc with two mirrored curves. This geometry will fit a sweep2 or networksrf.

If I use Sweep2 to fill this area I get this where the surface deviates from the curve significantly and is not symmetrical on both sides:

But, if I use networksrf, the surface does mach the curve (albeit with more complexity).

What is the preferred solution here?

Which “Cross-section curve options” are you using? Make sure you select “Do not simplify” or “Refit within” if you want a tighter tolerance between the curve and the surface than the one set in the file.

Let me follow on with this problem based on the original.

I am trying to add the three closed solids. Boolean Union simply fails.

I can do a splits with the Red and Purple and join them up.

However, I cannot split the blue with the red (but can split the red with the blue). And it will not do a boolean union.

I though this might be due to the coincident edge at the top of the blue and bottom of the red. However, even if I offset the blue away from the edge, I get the same problem.

I am curious why this is happening.

Problem Addition.3dm (3.5 MB)

My opinion. Please pass up boolean when too many get a tight edges
I hope the file can help even if there is no description.
Problem Addition.3dm (4.7 MB)


Your number one problem is your reliance on Booleans

If you would learn to use Rhinos surface modeling tools you would get the job done 10 times faster and end up with a lot better quality model.

In the enclosed file the blue and green surfaces are the only
part of your model that is relevant to what you are trying to accomplish. After reducing the problem to just what is relevant, it becomes simple to complete the model. Just use trim command to trim away the parts of the blue and green that you don’t need and then join everything. After that you can fill the one hole with a simple sweep2 surface.

No Problem.3dm (453.4 KB)

I was not relying on booleans here. I pointed out that neither boolean nor surface operations work works. I have presented a simplified version of the larger problem here. What I have in this particular case is that I can split A with B and B with A but I cannot do the two operations in sequence. After I split A with B, I cannot split B with A.


You summed up the cause of your problem with this statement in your first post:[quote=“bigjimslade, post:4, topic:41388”]
I am trying to add the three closed solids.

Why waste time creating all the extra surfaces to make 3 closed solids?
The probability of success will increase if you reduce the problem to just the surfaces that intersect. Adding more surfaces that aren’t needed will only increase the chance of failure.

With a revolution, you get the solid to begin with. I shared the solid version because it’s easier to create the services from them than to create “solids” from the surfaces.

As I said before, If I make these into two intersecting surfaces, I can split either using the other BUT after doing a split, I cannot split the other using the remainder.

That is the puzzlement.

@bigjimslade Is this what you are trying to achieve? Problem Addition DC1.3dm (528.6 KB)

I created it using Trim to eliminate the internal surfaces. To trim several surfaces I needed to use Intersect to create a curve to trim with. I also needed to explode one of the polysurfaces and trim the individual surfaces.

That’s is it. The reason for asking why is that area is actually more complicated.

This is a a work in progress so it will be more complicated than this. I was just testing with a simple block first and finding that did not work. This does not work either—obvious if the simple does not work, the complicated will not either.

If you are thinking clearly you will realize you don’t need a solid to begin with.

The reason is Split is the wrong command, Trim is the correct command to use.
If you think about it the answer to your question is sort of obvious, After you split one of the objects its no longer the same object. If you want to do what you suggest then make a copy of the first object you split to use to split the second.

But using Split command here is bad technique. Your model is an example of why its bad to use Split and good to use Trim command. Either one might succeed for doing the cutting, but if you want the objects to join after the cutting the Trim command is more reliable to make sure that happens correctly. The short explanation of why, is the Trim command calculates both the new boundaries at the same time and that pretty much guarantees the 2 polysurfaces will have matching boundaries that join up. If you do 2 separate split commands there is no guarantee that the edges will match up and join, particularly at the end points of the edges (vertices).

Doing 2 split operations on intersecting polysurfaces is a good way to end up with failed joins or the tiny orphaned edge loops at the end of edges that are supposed to match.