Bad Surfaces

[miano]

This is a problem I encounter frequently:

In the attached I have a polysurface with Four three-sided holes. Two have been filled in.

I creates two surfaces to fill the holes (red) using sweep 2. The objective is to have the two of the sides tangent.

I have tried edgesrf and networksrf for this type of problem and it never comes close.

I have tried cutting a triangular piece out to get a 4-sided hole but have never gotten a shape the looks close,

Problem Bad Surface.3dm.zip (779.8 KB)

[miano]

With a lot of trial and error one can get a sequence of sweep2 and matchsrf that will work.

[Tom_P]

i think the first step to do is, to improve the quality of the curvature of the surrounding faces (and try to modell those surfaces with the minimum amount of CVs)
Especially avoid those changes of curvature (side).

(showing curvature graph)

[pascal (Pascal Golay)]

Hello - I would back up a few steps and think more about how to put this together more cleanly to begin with - what you have is, it seems to me, an overly complex set of patches that (probably) do not really reflect the design very clearly - what is this? Is there a reference image or something that you are wortking from? My guess is there is a lot of scope for cleanup here that will ultimately make this easier. Just a guess.

-Pascal

[miano]

I am working off of blueprints.

Here’s what happens. This comes from a ship hull. For most of the length of the hull things are pretty regular. Large areas are simple edgesrf. Then you get to the propeller shaft exist where we have several places where radii go to zero where surfaces meet at angles.

The only way I know of to do an angle like that is to have the surface end at the angle.

The shaft bossing is defined as three arcs, some of which have lines connecting them.

Below the shaft bossing the shape give a natural triangle. I get around that but cutting out part of the surface forward to create a four-sided opening (where I don’t need a taper to zero radius).

Sweep2 is the only Rhino construct what does not want to bulges toward the center of the curve. If I do this as a networksrf in one four-sided piece I get:

The reason I have so many pieces here is so that the surface edges line up. The diagonal orange here could be combined with the gray to the left but It causes matchsrf horors and even more bad surfaces on joints. Instead, the orange slivers match the sides of the problematic areas that have to taper to zero radius.

The root problem from which all others flow is places where there needs to be a radius taper to zero at corners.

[pascal (Pascal Golay)]

I’d be inclined to try to stick it togther more like this

then do some trimming.

-Pascal

[Tom_P]

what s the result of those other / older topic ?

@miano

tangent.

It

FEATURE REQUEST: MatchSrf

I wish the MultipleMatch option on the MatchSrf command would permit selecting for the continuity for each edge rather than forcing the same for all edges.

@miano

.

T