Hi Mark, so what I'm talking about is right about in this area:



Right about there begins a significant mutation in the shape.

That then becomes a situation where you are actually trying to loft too much at once rather than not enough at once... I know it's tricky to find the right balance.

When you try to make a protrusion like that all in one go it will be very difficult to avoid awkward shaping.



Basically what happens is that MoI tries to match the whole length of each profile evenly with the next one.

It's not smart enough to figure out all on its own that different regions of your profiles should be matched with only some individual portions stretching out.


To try and describe that a bit better, in your situation there it looks like you want these portions to be mapped more closely to one another:




But then have these parts at the bottom kind of stretch out separately from the above parts:




Loft does not understand these kinds of sub-region matching automatically. You can help it by splitting the profiles into segments so that the segments will control the matching but when you have such a major shift in shaping that more calls out for a separate piece there and attach it to the main one by fillet or blend.

- Michael

Mark,

Warship hulls using NURBS is both my nemesis and sweet spot. I do not have time right now to go into any detail (but will soon) about how to help you.

I think that you are making things very hard for yourself by having too many curves (stations that is), that are irregular and with many points. Also, as said before (I zoomed past the thread), it might be counterproductive to generate the entire hull in one go.

British cruiser hulls nearly always display a knuckle at the bow. This you can add later and therefore disregard the part above the knuckle for the hull generation proper.

In my experience, it is better to work with longitudinal curves (waterlines) than with vertical sections (stations) and to use only as few as absolutely needed to generate the shape you want. Using curves with the absolute minimum number of control points also pays handsomely. 5 points are all it takes to control a simple curve, generally: begin, end, largest deviation from straight line and one point either side of it to control the curvature.

The stern is a more difficult area, but one that can be tackled rather painlessly using a combination of network curves, lofts, blends and sweeps depending on the case.

I will get back to this when I have time to work out an example.

Hi Michael,

Many thanks for your continued efforts with this. I've decided to go with my networked hull for this model. It has a hollow which I didn't notice before which is bothering me now but much additional work has been done on it (plate extrusions etc) which I've decided I don't have time to redo.

This has been a very worthwhile exercise with some water still to flow under the bridge (ship?). I have another one to do which I will attempt with a loft and trim mindset from the beginning.

Jean-Paul, I eagerly await your thoughts on this. I have already referred to several of your threads for ideas.

>In my experience, it is better to work with longitudinal curves
>(waterlines) than with vertical sections (stations) and to use
>only as few as absolutely needed to generate the shape you
>want.

Unfortunately, longitudinal curves are almost never given in the plans I have so I have had to work from stations.

I couldn't resist attaching latest progress render and a two minute scene render (with some odd coloured water!).

Cheers

Hi Mark, well your current renders there are looking great!

- Michael

I wanted to fool around with this a bit. I talked to a guy who told me that using the waterlines as opposed to the stations is how they build hulls. The stations are more a way to get your waterlines. So I took your loose loft of the stations that created your shape fairly close to what you want, then I created a direction array of a few lines out to the side and projected them onto the side of the loft. Then I ran rebuild curve on those new waterlines and used about 20 points to get curves with no deviation. I then lofted with "normal" the waterlines and got a good points structure that could be manipulated a bit easier to adjust areas. I could do this while the loft was still active in history and just use the points from the input curves.

I think I could go back to Station manipulation by reversing the process, but it is much harder to manipulate the overall shape by stations.




Any further input from you Mark, and JP is appreciated.

Hi Burr,

Wow, that looks great. Thanks so much for following this through. I wonder if that is the technique that Jean-Paul uses?

I need to have a go at this today. I think I understand the technique and will get back to you with results. I really hope I can duplicate what you have done there as this is looking like the method for hull generation.

What a great forum this is. Thanks again Burr :)

Mark

There is a tutorial in the Rhino website from the guy that did the Boat that is on their main page. He has a model that you can download to look at his tutorial, and in his model this is what he did.

I revisited it after I posted and the waterline curves, rebuilt at 20 points, actually do deviate a hair from the original curve. Really only at the bow and stern. More points just means harder to work with, but this does give a different angle to adjust the shape in some areas. It may be small enough to still work it this way.

Good luck.