Need Help With Aluminum Lofting

[jdnelson97]

I need someone to explain how I can come up with an accurate Aluminum Flat pattern for a Hull plate and Side plates. I am currently using Rhino and have the keel line, Chine and Shear line drawn out. But When I create a surface they are curved but are not accurate.

Thanks

[gonzo]

How do you know they are not accurate?

[jdnelson97]

I have drafted a few boats with rhino and have sent the flat patterns to have them cnc cut. Most of the parts came out accurate except for the hull plate and side Plates. TheY were off at the chine edge and shear line.

[DCockey]

How are you creating the surfaces? Which command(s) are you using?

[DCockey]

How far were they off?

[daiquiri]

Could also be a matter of the precision of the CNC cutter. I have seen 6 meters long CNC cut parts with 10+ mm error in length. If not corrected, that's enough to create a miss-alignment of the chine edges.
Try to measure the cut parts and confront the measures with the CAD drawing of the developed surfaces.

[philSweet]

Is your problem only that the hull plates bow away from straight sided frames when you assemble them, or is there more going on here? I don't use Rhino, so I'm only wondering about your methods, not the software. If the frames were correctly shaped, would this solve the bad fit?

One cannot, as a rule, draft arbitrary keel and chine curves and get a panel to fit them, although some software will do their damndest to try to. Some just use a least squares fit to find the "best" approximation and leave it up to you to do the developability check.
You might also try doubling the number of stations on the linesplan and comparing the new plate output to the old. That might indicate a precision issue.

[DCockey]

You need to start with a developable surface.

Create the surface using DevSrf. With Rhino 4 it needs to be downloaded from from the McNeel site as a plug-in http://wiki.mcneel.com/labs/devsrf It's included with Rhino 5.

You will almost certainly need to extend the curves beyond the hull to get a surface with DevSrf over the entire area desired. DevSrf has an option, "extension", which does this though you don't have control over the shape of the curves. Or you can extend the curves before creating the surface. The shape of the extensions will affect the shape of the surfaces. The shape of the side surfaces in turn controls the shape of the stem where the side surfaces intersect. Generally some interation is needed to find the curve extensions which yield the desired surface shape. The surface will need to be trimmed back to the desired size after it's created. Side panels are trimmed at the centerplane or "rabbet" line.

Loft is another command which can be used to create a developed surface when the style is set to "Developable" in the pop-up window. It works but can lead to problems. It tries to create developable surfaces abuting the entire extents of the curves which would seem to be a good thing. But to do so it uses "triangulation" to fill in the gaps. Additional surfaces with zero length on one side are created which are developable but which have zero radius of curvature at the zero length side. This can create problems in fabrication. Also the surfaces would generally be considered as not fair. Extending the curves eliminates this problem. Loft also works with extending the curves though it's not as obvious with Loft as with DevSrf when the curves need to be extended.

[DCockey]

A common misconception is the idea that the sections of a V-bottom hull with developed surfaces are made up of straight lines. This is only true in a few cases, usually the sections will be curved. A surface with all sections as straight lines may not be a developed surface.

[DCockey]

Originally Posted by philSweet

......
One cannot, as a rule, draft arbitrary keel and chine curves and get a panel to fit them, although some software will do their damndest to try to. Some just use a least squares fit to find the "best" approximation and leave it up to you to do the developability check.
You might also try doubling the number of stations on the linesplan and comparing the new plate output to the old. That might indicate a precision issue.

By extending the keel, chine and/or sheer curves panels can be fit to many sets of curves, though not all. That method does however determine the shapes of frames, stem, transom, etc though the choice of shape of the extension curves provides some control.

[DCockey]

It can be tempting with Rhino to use Squish or the older Smash rather than UnRollSrf to flatten panels and determine their shape for fabrication. Squish and Smash will flatten surfaces which are not developable and so it would seem simplify the process. But the resulting shapes won't simply roll up to the desired shape without some stretching and/or shrinking. UnRollSrf works only with developable surfaces and a surface flattened with it should roll up to the desired shape without stretching or shrinking.

[philSweet]

One other simple way to check on the robustness of the software. Take the entire hull and rotate nose up 45 degrees and repeat panel development. Ditto nose down. The panels should come out the same, but some methods lose accuracy depending on the projected curvature. Jerking the software around like this can allow you to operate within acceptable bounds.

[jdnelson97]

Thanks to all for the input. I believe I am back on track. Thanks for the PDF Phil!

[jdnelson97]

The info that Phil shared will help me to be allot more accurate when creating parts nesting. Here is a few recent projects that I have completed. The issues have been fairly minor, but I strive to be a perfectionist!

Thanks again to all!!

[DCockey]

The user has considerable control over the accuracy in Rhino with the settings in the "Units" portion of the "Rhino Options" page. You can get to the page by the DocumentProperties (Properties on Edit menu) or Options (Options on Tools menu). Then expand the Units section on the page which opens.

Frequently a smaller tolerance is helpful for the tolerances. For small boats I generally use:
Absolute Tolerance .001 feet or .0001 feet
Relative Tolerance .1 % or .01 %
Angle Tolerance .1 or .01 degree