Modeling boats in Freeship-The techniques of Using the sofware

I presume you have already tried to take the check mark out of the corner box in the point properties pop up box. I suspect that the way the edges were extruded here is the culprit...not allowing the corner to be deselected. I would save it then save a copy and start by deleting the point and see what happens. Perhaps you can recreate it without having the corner being hard.

 

Frank
The points that define the shape are a control net, not necessarily on the hull. If all faces were flat panels you would have a faceted hull and all control points would actually be on the hull surface. However generally the points of control net are not actually on the surface - the corner point at three or more creases will be on the hull. The actual line of the surface is shown as a thin dark line that is usually a slight distance away from the control points because it is smoothed.

The image below shows a similar situation to what you have. With the point at the stern I split the edge where the edge was close to the actual hull shape and extended in different directions there. The Gaussian curvature shows that this produces a less tortured plate compared to the bow where I have extended from a point on the existing net which is further away from the line of the hull.

In practice you would be making these sections from different pieces of plate or sheet and each would have their own control net. You simply adjust the control net for each piece so the true outline for each are within your cutting tolerance.

There other ways to avoid the severe distortion but the important realisation is that the points and lines of the control net are not necessarily on the hull.

Rick W

 

But for this case, the important realization is that the hull surface actually jumps out to touch the net at this point to form a singular, peaked location:

The nice original curve of the sheer is destroyed and the actual control curve is now split in two. The problem workaround is to move that point back exactly to where the actual sheer curve was before (if you can remember or had it recorded – say by export and reimport to markers) AND then to add a point on each side and fairly near to this singularity in order to try and get some curvature back close to the peaked location so that it appears as if the sheer is continuous like it was before. It unfortunately doesn’t help that the curvature analysis now does not work thru this point. (Hopefully martijn will improve the tool for delftship so that it doesn’t unfortunately go to 0 at the end of all curves - then it would be useful for these and other situations)

While the abrupt change is especially obvious with crease intersections, whenever a point is added to an edge curve, the original curve is changed – even when it is ‘locked’, its influence location is locked, not the affected real curve. And where the curvature is great, the affect is large. If one is trying to keep the original hull sheer condition, even the right side point addition in Rick’s example causes the sheer to be changed from where it was originallyto be exactly touching that intersection. That it doesn’t add much deformation, may or may not be an issue( but it will be for the in-plane sheer curve!) – but it does make a change that might be an issue – especially if one is trying different cabin tops and expecting the hull to remain the same shape!

When freeship/delftship develops a curvature analysis tool that will work through these typical 3 edged situations, it will change from a schematic tool to a much more useful modeling tool.

Here’s another example of exactly the same thing: I wish to have some way, other than just my eye, to make a fairish curve of the end of the strakes of the Oseberg Ship (highlighted points and 'lines'). Or I wish that I can keep the hull sheer strake in a fairish curve when I add the gunwale strakes. Right now it is just discontinuity after discontinuity - right now, I can only do it by eye:

 

Hard chined boats in Freeship / Delftship

It is important that you connect the points in the right way so you not get concave waterlines in front. The bottom is still developable, even more natural.

JS

www.sassdesign.net

 

JS , I found that also , I have thought of adding a point as an apex for a guide in development of the bow . This would be out in space , with a line going to 2 chine points .

 

Do NOT forget...adding a "point" will throw off Freeship/delftship more than you really want. The program is Seriously optimized for a 4 point face...anything else tends to give it fits or at least heartburn. If you can...move a point of a 4 point face to get where you need to be rather than add a point and make it an uneven point numbered face. If necessary...add enough points to create another 4 point face...this will help but might not be the optimal.

 

[rant] And to think...By Eye used to be the ONLY way to build a boat...it is now relegated to an apology for the perceived inadequacies of a program. If it LOOKS right in the software...chances are it will come out within .0025" of what you want...how close do you really really need it to be when a plane stroke is enough to get you off by .005"?

Somehow I think the search for the CAD "perfection" has gone a bit too far. When you start talking about .0001" of an error as something significant in wood...you have lost the spirit of the thing and given over to the mechanics and sterility of the drawing program and not the warmth and tradition of building in wood....after all...even the sheets of peeled and laminated wood veneers called plywood are unstable and subject to the environment. And when talking plywood... .00520833333333' (1/16") is a SMALL measurement in Stitch and glue and the Maximum variance in the old measurement of offsets of a traditional planked hull was (1,1,5+)(inches,1/8's,1/6ths) and was only +/-1/16th of an inch when absolutely necessary. It is very easy to cover an 1/8" variance in ply with the Wonder of wonders...Epoxy...or even more with a sloppy fillet. Do we really NEED to worry about the precise fitting of joints and panels to the hundredths of an inch or is it just stick in the mud tradition applied to an ever diminishing traditional thought line of building which advocates precision in design but "give" in building techniques because someone uses a plane to shape the planks? I use a plane...I use a cordless saw...a drill ...a jig saw, ...a circular Saw.... a sawsall,...a grinder...whatever gets the job done,,,but there are those of us who scorn others if they even THINK of using these techniques. [/rant]

Sorry...didn't mean to rain on your parade but this got me a-huffin' and a-puffin'!...I'll soon be back to using software to modeling hulls to the .0001 foot.
/100.

 

What drives me crazy is when I export to my CAD program and it tells me the surfaces don't intersect. Then I check the numbers and it is off by .00001, grrrrrr!

 

Contrary to your understanding, I am attempting to positively contribute to your thread on freeship techniques. The realization being missed is that inaccuracies are about ½” at the end of some strakes and up to 2” on the gunwale intersection.

The issue I think illuminating, is that the peaked point situation (that Frank raised) happens everywhere that an internal corner point is placed or whenever 3 creases or edges come together. As well, it can happen in the surface and/or an edge at single or multiple locations along a previously ‘faired’ curve edge.

To try and work through this again with a simple situation, here is a 2 panel boat defined by only 3 points along its length – endpoints and some mid point. Although this is a crude example, let’s assume that I designed it and love it as it is – and have resolved lots of measurements both dynamic and static to get to it, or that I already built it this way, or that I just bought it and it is now sitting in the yard. In other words, I spent big bucks, or big time, or big emotion getting to this hull and it is way too late to change it now:

(To re-emphasize – this is a general example; it could be a stem, gunwale, or keel where a strake or panels end, or a multi faceted cabin, or a deck panel intersecting the cabin side or many other situations.):
fig 01


The first thing to note is that the gunwale is a 'fair' curve. The second feature to note is that point 2 is separated from the gunwale by a significant distance.

So after acquiring this hull (by whatever means: imagination, money or build), I wish to add a most simple cabin top. As this is to demonstrate the general issue that possibly could be an immersed situation below some waterline, I will need to join the cabin to the hull.
I show the cabin just floating above the hull by 1 foot so one can see the intention:
Fig 02


And now here it is unjoined, but dropped 1 foot, approximately in place:
Fig 03

Note that in a typical case the cabin would be modeled directly and not imported to show two separate pieces. This approach is purely to attempt to clearly demonstrate the issues of what happens next. Note that the hull is unchanged and fair and that the cabin tent is also.

So now there is a weird choice:
1) does one join the two pieces so that the original hull point 2 is retained? That’s the point that truly has defined the hull from now, so shouldn’t that be paramount? But here’s what happens if the cabin joins to the previously good point 2:
Fig 04

This contortion is obviously not a straight forward approach.

2) Or alternatively should the cabin corner near to point 2 be retained as that is where the gunwale sort of ends up?
We read that corner points are right on the surface and corner, so it does make some sense. The first problem here is that one can’t ‘snap’ the cabin corner to the gunwale curve like in other programs, so one has to do it by eye and zoom in so that its location is close enough for our desired tolerances. A little slow and cumbersome as one has to do it in 2 views, but no big deal:
Fig 05


This approach seems works in some fashion, but now the surface is radically dished and unfortunately the rear half of the gunwale curve has been changed into a straight line. And furthermore, the gunwale’s original location is lost.

So to now change the rear gunwale straight line back into a curve, one has to at least add a point. And to take the dish out of the front part of the topside(s) panel, one has to add a point to the front half of the chine curve and join these 2 frontal points up with a ‘normal edge’:
Fig 06

But now the added rear point dishes the rear of the panel. And the simple point added with no moves at the front chine changes the lower chine curve at the front. The change at the chine of this 12 foot boat is on the order of 2 full inches. And added to this situation, the original chine curve location is lost as well and I think we can foresee that a rear chine point must be added as well.

**

I hope the above demonstrates the cascading series of adjustments that must be made when one adds a single hard point in the edge of a curved surface. Highly curves surfaces or edges are much more affected than those with minimal curvature.

So some considerations come from this:

• one cannot retain original hull shape when additions are made.
• New points and normal edges must be added to approximate the original shape
• Any new points make changes in edges and in surfaces that in turn may require reaction.
• Changes can happen in edges that are adjacent to edges that are changed

Some procedures are required to assist getting back to the original shape: First, save the original unadorned hull shape as an inviolate, separate file. Then there are some options:

1) save the linesplan and use it as a background image in the 3 views. After the cabin is added, add and shift points around as necessary to match the background. This approach is exact as the graininess of the lineplan bitmap.
2) Export offsets (only when model is in ‘highest’ resolution) to file, strip off everything except control curves, and read back as markers.
3) Another is to export original as part, read in, change colour, lock all points and remove from hydros. I haven’t tried it, but feels cumbersome and confusing.

Using approach 2, here’s what the markers look like just as the cabin is added and before any steps are taken to try and match things up (2 pics ago-fig 05). The markers coincide with the chine, the surface is dished, but the gunwale is close at the front but the stern gunwale portion is a straight line that misses the original marker curve by probably 4 to 6 inches (I forgot to measure it):
Fig 07


You can see that the markers seem to coincide with the gunwale, but the actual case is that I cannot get better than ¼” error with solely adding (and moving around to 'best' fit) the bow 'quarter' point and the stern 'quarter' point. You can plainly see how out of whack that the chine is before adjustment with the simple addition of the chine bow point. Before adjusting, this is out about 2”.
Fig 08



summary

• Adding a hard point can have significant effects on areas of high curvature
• it is not always a trivial exercise to account for the addition – in some cases many new points will be required
• one compensating aid is the use of output/input markering.

​

-mick

 

Mick. , that is what sends me back to pencil and paper . On paper I can put something where I want it and it stays there . But getting quick hydros is good in Freeships . I dont use it as an illustration tool , but for playing with ideas .

 

Why are you using a separate panel to create your top instead of extruding the edge and shaping from there? You would then not have the corner issues because your are adding a surface along the same edge and it would be pulled by the same point along the edge or another point with the same curvature properties would be created if you extruded up then in. You can even combine both at the same time to get the angle of the tent. It might not work for everything but it will definitely work for your example. When adding .parts the expectation is to get it to the same or very close to the modeled surface by adding as many points to the part as needed but not actually joining the hull. Any surface work should be developed by extruding the hull. You point out the exact problem because the part is edged with hard points that often don't want to change...the trick is to add soft points to make the lines conform to faces. Trying to join a hard point to a soft face will only distort the mesh in a very nasty way...as you point out. It is a limitation but it can be worked around for the most part. You try to model the exterior surfaces by using the exterior surfaces and points and extruding. Adding interior stuff can be done using .parts or by separating extrusions and moving into place and shaping. Always extrude up and in if it will affect the hydrodynamics with leak points. A leak point...even inside a hull will throw things off.

 

Probably because where the edge of the top meets the sheerline a hard point is created. This happens automatically whenever 3 crease edges meet in a point.

 

you have to do the deck and bottom separately. I extrude the thickness of the
hull and then finish the deck , and get rid of the face between . Works out nice for floating the deck structure . Actually breaking the boat down to components works well , but is time consuming. It does not work with areas below the waterline of course .

 

Martijn...are you open to suggestions here or should I email the couple that I have direct?

 

When I near 'completion' of the hull shape I rely heavily on the resistance performance curve. This indicates if the shape improves or worsen.

If you select the view mode to developability check you can see if the shape is smooth or out.