TurboCAD Forums

The Ultimate Resource for TurboCAD Knowledge

Register
 
The purpose of these forums is to discuss TurboCAD.  Please keep the discussion of other products to compatibility issues or how the other software’s can be used with TurboCAD.


Irregular Lofting on a Boat Hull
Read 1442 times
* February 17, 2019, 09:03:18 PM
#50
I don't agree with you about deriving the loft from frame stations.   "As built" frames are spaced to attach physical skin to the framework in the most practical manner, using them as profiles to loft often creates surfaces that don't follow intended flows.  It's almost inevitable that more profiles (which don't fit the station spacing anyway) are added to subdivide the loft, more profiles often make the loft less smooth.  Better to loft through as few profiles as actually needed, and to vary the distance between them to modify local curvature or bulge rather than thinking that the loft will obligingly flow as intended through equidistant profiles, an experience like sighting unicorns in my experience.  Hulls and fuselages taper.  THEN derive the rib frame or bulkhead profiles from the hull's profile at their stations.   TC's section and intersection tools, and the drafting palette, give you a lot of ways to derive those contours from a minimal loft more likely to be fair and smooth than a hull that fits equidistant profiles exactly at those stations and evokes Bibendum the Michelin man between them.

Logged


* February 18, 2019, 01:50:30 AM
#51
I also agree with Murray's post but from a practical viewpoint. Like Murray says that using too many profiles can lead to distortion of the loft and thats exactly what Andy H picked up on with one of my earlier posts. Point is Andy noticed that the loft was stressed, which meant that the loft had irregular lines going in many different directions and I had trouble controlling the general loft shape. At the time in that post I had used too many profile splines to try and get a nice loft. After taking Andys advice I achieved a good loft shape when I reduced the total number of profiles.

Logged
Daz
TCW V21, 2016-2018 PP, Animation Lab V5, TurboPDF V2-V3 & LightWorks Rendering Engine mostly.


February 18, 2019, 02:48:16 AM
#52
If you have an accurate table of offsets, for a known fair hull, or for that matter an aircraft's fuselage, wing et cetera, then it is reasonable to use these to derive the loft.

However, if you are designing from scratch or simply guessing, then it is easier to use fewer profiles to get fair curves and thus a good loft. 
These curves and surfaces can be checked in suitable software. 
Unfortunately, TurboCAD is lagging a bit behind the competition, in this respect.

Regards Tim

Logged
You can design without engineering, but you cannot engineer without design.
Using Win 10 with Designer 2017 and TurboCAD Pro. Plat. 2016/2017/2018 + Lightworks (64-bit versions) + AnimationLab.


* February 18, 2019, 03:42:35 AM
#53
Tim, I disagree that it's reasonable to use a table of offsets at station points to loft through, especially at the bow and stem of vessels where they taper, for the reasons that I outlined above.  In most use cases, TC won't deliver a satisfactory result by doing that.  It's certainly possible for very sophisticated surface development programs like Rhino, which can define degree 32 spline surfaces, or Maya, or Alias Studio as used by car corporations, or for DelftShip or its predecessor Freeship, but those last two create subdivision surface hulls, not NURB, which don't interact well with associative or Boolean functions in TC or CAD generally.   Most of us aren't using exotics, we're using generic CAD lofting functions like those provided by ACIS, Parasolid or OpenCascade, which will generally extrapolate degree 3 (maybe 4, if I haven't remembered the relationship between degree and order, one is greater by one than the other) spline surfaces through the V direction of a loft, which doesn't give that sort of control between stations.  You could work around to some extent by changing the direction of the loft: if you dig into TC's spline functions, you'll find that you can draw curves of up to degree 9, which have a higher tension, resistance to kinking than curves of lower degree/order.  Taking that into account, you could draw your hull using longitudinal profiles that could be smoother in the bow-to-stern direction.  It might be easier to control a loft in the keel-to-gunwhales direction using longitudinal profiles.  Tables of offsets are somewhat like NACA/NASA airfoil profiles in that they do give you definite points that the surface OUGHT to interpolate, but they don't take any account of the capacity of the tool that you're using to actually do that.  That's why I recommend PolyCAD for hull development: it has functions dedicated to that.         

Logged


February 27, 2019, 09:34:11 AM
#54
I also like playing with boat hull models but I do it slightly differently.  I have attached spline dividers which I add to the model at right angles to the cross sections on the centre line with the flat top part level with the highest point on the cross section that is "highest".  I then shift it up to a level abut twice that of the tallest cross section.

In each turn I move a cross section by exactly the same as I moved the dividers.  I then replot each cross section using the SEKE I where that cross section crosses a divider line.  I use either spline by control points or beziers (my preference) but I do the deck and the hull separately for only from the centreline outwards.  Ultimately when each cross section will be moved back down to original position, the deck and hull beziers are mirrored about the centreline to ensure symmetry.  Note to check after moving each back to its original height that the position along the hull is correct as replotting sometimes mean the new beziers is at PosX but that is simple to copy the PosX or the original cross section into the new bezier to get it in the right place..  When the hull sections are lofted because there is the same node in each section at contiguous positions. the lofts are "smooth".  A fag, yes, but worth it in the end.

I note the bow and stern sections are done separately which is wise as lofting of those can be troublesome.  I see also that the sections are not uniformly spaced and this can add to the uncertainty.  I think also the variations between the cross sections is to severe for good lofting.

As a demonstration of my methods,  I have attached a screenshot of my current work in progress of the Yorkshire coble "Eliza" before mirroring and lots of detail of masts, sails, transom and rudders.

Logged
regards
Colin Reid
TC2017 Pro Platinum 64 bit  + LW plugin on Win10 desktop and  Quad Intel i7 with 32GB RAM + 128GB  SSD and 2TB partitioned hard disk, NVidia 2GB video


* February 27, 2019, 12:59:43 PM
#55
Thank you Colin for your helpful insights ;D

Logged
Daz
TCW V21, 2016-2018 PP, Animation Lab V5, TurboPDF V2-V3 & LightWorks Rendering Engine mostly.


* February 27, 2019, 08:37:15 PM
#56
Colin I presume that you use your spline dividers in left or right view as an aid to designing the spline profiles. Is that correct?

Logged
Daz
TCW V21, 2016-2018 PP, Animation Lab V5, TurboPDF V2-V3 & LightWorks Rendering Engine mostly.