Fairing hull surfaces and resolving edge anomolies

[gileshg (Giles)]

Hi All,

I’m struggling with a surface that I’ve lofted throught a series of mesh_intersects.

The lofted surface has been edited using only the surface control points and not the original curves and history, for time saving really. The result had a few anomolies that I corrected by dragging control points around. However, I still have a surface which has a couple of holes at the point where my original curves started to sweep around the ends to close the loft. The surface is also quite tight (by this i mean that it’s highly sensitive to manipulation, giving a fairly unfair surface. Is it possible to make the lofting a bit “looser” at this stage?

Lastly, how can I get the ends to be represented as a fair nose/tail and not segmented as seen in the last image?

many thanks in advance for your thoughts!

[Stratosfear]

Your form is too complex to do with a single surface. Overbuild trim and fillet will produce better results.

[gileshg (Giles)]

Thanks Stratosfear. I feared (sorry) that might have been the case. I’ll make some more surfaces and have a go at trimming.

[Stratosfear]

Nothing to fear. It’s easy with some practice.

[gileshg (Giles)]

are you splitting your sections into manageable segments and lofting through each series to create a collection of surfaces?

lovely looking kayak. Has it been made yet?

[Stratosfear]

I rough out the form with lofts. Then I’ll split it up to make bezier surface patches for all my primary surfaces. That dictates how many sections are needed to define the form. I also do some point editing with curvature graphs on. The hard part is always the trims and blends. I use the pipe command a lot to create the trims.

Last I know if it the molds were made. It was a long job started with an almost worthless scan and too many revisions.

[TomTom (Tom)]

Good design is also taking care about the details, and to control everything almost any shape must consist of multiple surface, which clearly determine the flow of form and reflection.
As @Stratosfear shown, you usually model from from big and low curved to small and highly curved surfaces. Always build theories, so all of your big main surfaces somehow match positionally (if possible). Then blending or filleting them always works well.

However ellipsoid shapes are always difficult to model because they always have strong curvature in multiple directions.

[gileshg (Giles)]

my job is also from a scan. I hadn’t thought of turnig on curvature graph for the control points on the surface. that’s a great idea.

It’s frustrating because it seems really close to a nice model, just lacking some fairness and closure

[gileshg (Giles)]

I’d love to be able to model it without the dfefects though!

[gileshg (Giles)]

@TomTom

However ellipsoid shapes are always difficult to model because they always have strong curvature in multiple directions.

Yes, I’d find it difficult to pick where to split this as there are few obvious changes in radius/form. I’ll try the stratosfear method.

[Jess]

Hi Giles,
you’ll have to avoid cases like this:

Rhino does not like it when the u- and v- direction of a NURBS surface are parallel. The bad shading indicates that there is a problem. If you try to offset this surface it will probably fail or result in a bad object.

[gileshg (Giles)]

Jess, that is exacly one of the problems. So I’ve tried extending the surface and untwisting the edge. It’s time consuming but I’m thinking that I can then trip back to the centerline…

I’ve also started playing with multiple surfaces. Hopefully that will yield better results.

[heatherogden361]

Jess, that is exacly one of the problems. So I’ve tried extending the surface and untwisting the edge. It’s time consuming but I’m thinking that I can then trip back to the centerline…

I’ve also started playing with multiple surfaces. Hopefully that will yield better results.
thanks!