Rhino unrolled surface deviation

Hi again all,

As a follow up from my last post about the new 7.5m design i completed (nearly completed)

I am flattening down the parts, most of which, well all but two parts when unrolled do it and in the command line say "within tolerance" which is set to 0.1mm, apart from the bottom and topsides, both of which are coming in at 0.15% which i am told is too much deviation for the final model and it should be under 0.05% to ensure everything fits correctly during the build.

My questions are related to how others go about "lofting" surfaces. There are several ways i have tried, using the following commands, loft, sweep 2 rails, devloft, devsrf (plugin).

The one that i get the best results with is the sweep 2 rails one, but still cant seem to hit that 0.05% mark, the problem areas are where the twists are in the plates if you like, coming in to the bow.

Any ideas or tricks would be much appreciated.

Thanks
Robert

 

Hi Robert,
Even if it strays a bit, but it will be very very little, from the original shapes, avoid curved strokes in cross sections and trace them using straight line segments.
Use those segments for the "soft" command. That will create ruled surfaces that are easily developable.
In my opinion, a deviation in dimensions of 0.15% is very acceptable. I don't think more precision is needed.

 

I understand that you may be eager to move forward and start with information about construction processes and drawings. BUT, make sure, before starting this stage, that your boat is correct from all points of view of naval architecture and that it complies with all the regulations that apply to it. The famous spiral of the design can lead you to the situation of having to change the dimensions, the shapes of your boat or the general layout of your boat.

 

Yes i most certainly will, don't intend to build this till over next winter, so have a long time from now until then to make certain that everything is 100% ok. Thankyou for all your help on it, truly is much appreciated.

 

Looking at the pictures you have provided in the other thread, it is not possible to see the hull shape, and wether the actual panels de facto are developable. Now, when you create a 3D panel from two or three curves, the ruling process may give different answers depending on the order, in which you select the leading "rails". This occurs when there are more than one ruling tangent possible at various positions between the "rails". I'm not familiar with the Rhino software as such, but I've seen this occurring with other CAD systems.

Note the functional differences between "lofting", "sweeping" and "ruling"! In general, for single-curvature (ie unfoldable) shapes, you must use "ruling", otherwise there is an increased risk that you create double curvature.

The question arising from your pictures is if you have created the frames and transverse stiffeners before you ruled the bottom surface, and if yes, do you use the frame shapes (whatever they are) as "clamping" restrictions during the shaping of the panel. I'm asking this, because a very common mistake is to assume straight lines for the transverse members; for a correct multiconic, developable surface, the transverse boundaries between fi bottom and frame are cambered. With the bottom and side shapes seen in your pics, applying the wrong transverse shape would result in a slight double curvature in the skin. As TANSL notes, the amount of 0.15 % might be allowable, but it will result in unfair surfaces and unwanted foul language in the workshop!

 

Hi, thanks for your reply,
I started with a lines plan, then modelled and ruled the bottom first, then the chines, then the topsides, the frames were then put in after that. I have used the analysecurvature command and checked the curvature of the plates and they look good, that's why i'm struggling to see where this deviation is, i will be trying again tonight, maybe will break it down and start again, try to rebuild the surface and see if i can get those two to come out slightly better, might not acheive the 0.05 but hoping i will get closer to it.

 

A few tips:
-When you test the surface creating tools (here "ruling"), don't delete the original surface when you try the next option, just use a different colour for the new one. That will reveal what differences they create and where.
-It seems easier to produce truly developable surfaces when the directors ("railings") have monotonically changing radii in both Y and Z when moving from one end to the other; usually reducing bending radius when travelling aft-to-forwards.
-Creating a 3D surface by ruling between two curves, say chine and centerline leaves the third edge (say transom) undefined. If, fi, the transom is selected with the chine to represent one director and the centerline (keel plus stem) is the other, you get a different skin shape than when you select transom plus centerline as one and the chine as the other. A third shape (with a straight transom edge) will emerge when the bottom is ruled only between center line and chine.
-If the keel line is straight from transom and fw, there will be a limiting ruling from the fw end of the straight line to the aft corner of the chine. The shape of the triangular surface between straight keel, limiting ruling and transom will be determined only by the curvature of the transom's bottom edge.
-All lines and curves in a CAD model have a direction, some operations depend on direction for the calculation process, which may generate differing ruling outcomes from two identical directors.

 

Have had another few goes at it, best i've got is 0.11%, the plates up to midships are fine, no issues whatsoever, its when they start turning inwards to the bow, i split them up and unrolled them by section and it's only the bow sections that are coming out as 0.15 ish out, i've tried using the curvature analysis to check the curvature and look for any possible issues, but cant find any. Will maybe just have to be the way it is. It is saying the plates are larger by 0.1...% so whatever it is we can trim to fit when building I suppose.

 

Depending on what material you intend to use, the deviations shown will cause quite high stress levels. Using the E-modulus for aluminium (~70000 N/mm2) and stretching or shrinking 0,15 % will result in a stress value of 105 N/mm2, which is about 62 % of the maximum allowable for "normal" hull plating (DNV 15-m rules). In my book that is quite a lot. If steel, it's about three times as high.

One problem is that when you unfold a convex/convex surface (like a bulb), ie from 3D to 2D, there is tension along the edges and compression in the mid-zone. Going the other way, from 2D to 3D, there will be compression along the edges, which needs edge shrinking and/or mid-zone rolling with an English wheel or equivalent. In this case the edges need trimming. Your notion that the panel is defined as oversize tells us that the software is using the 2D to 3D procedure, which is the preferred way. For a convex/concave surface (like a saddle), it is the reverse; the edges are stretched when going 2D to 3D. It is important to know these differences in order to apply the correct procedure during fitting. Without access to your plans it is impossible to give more detailed advise, so good luck!