Creating frames, stringers, concentrated masses etc. with Freeship / Delftship

[daiquiri]

Hello,

This thread is the continuation of the discussion started here: Best Free or Low Cost Marine Design Software? (2011) , which regards the methods (if any) to create bulkheads, frames, stringers and other structural features of the hull with Freeship and/or Delftship.

The basic intention is to allow as many of these features to be created in Freeship/Delftship, in order to speed-up the design process by avoiding the need to export to other software like Rhino, Solidworks and similar.

Please feel free to contribute with your experience and tips and tricks (about these or other useful tasks) which could be of help for Freeship / Delftship users.

Cheers!

[daiquiri]

In the previous discussion an idea has emerged (Best Free or Low Cost Marine Design Software? (2011)), which goes like this (example for vertical frames):
- create vertical planes where you want the frames to be positioned
- place frames and hull in separate layers
- find the intersection of the two layers
- extrude the edges thus created, to form a frame or a bulkhead.

This is a simple example I've created to try this procedure:
Canoe-like boat.fbm
It is a Freeship .fbm file. I'd be very gratefull if someone checked it out and told me why doesn't it work as desired... Why cannot it find an intersection of vertical planes and a hull wall, since they are obviously intersecting?

Cheers!

[daiquiri]

Another question which I'd like to discuss is: how do you create concentrated masses with Freeship? Like the engine, tanks, deck items etc.
Cheers

[Olav]

daiquiri,

I downloaded your file and checked it out:

When I tried to intersect the layers in FREE!ship 3.2+ I continuously got an error message saying

Access violation at address 004B2D5B in module 'FreeShip.exe'. Read of address 465C3A48.

Seems to be a bug in the latest version of FREE!ship...

However, it was possible to export the "hull" and "frame" layers as a part and import it in both FREE!ship 3.12+ and DelftShip (which is necessary as older versions cannot read *.fbm files created in newer releases, but *.part files are upwardly compatible). Here it was no problem to create new points and lines on the hull shell layer.

I'm not quite sure if this is what you tried to do since it is not possible to extrude these new curves without some "tricks". You could adopt the coordinates of the new points and modify your "frame" layers manually or delete the hull surface to the frame location (to make the intersection lines an edge), add a new layer and extrude the edge and then go on with the next frame accordingly. The latter approach destroys the hull surface which would have to be saved before as a part and then re-imported.

If you intended to create new points on the "frame" layers: This is not possible, because both FREE!ship and DelftShip only insert new points on lines and edges that pass through the intersection plane. The outer edges of your "frame" layers are clearly outside the intersection.

Re concentrated masses: I never do this in FREE!ship or DelftShip, but my suggestion is to insert a new layer (as a *.part or a layer created with the "Insert Cylinder" option), make it a small rectangular shape and adjust the density of that layer in such a way that it represents that particular item. I would do this independently from a part that represents the outer dimensions of the engine, deck fitting or the like.

[daiquiri]

Olav, thank you for the reply. I just saw that Perm Stress has already shed a light to that problem, and I think I now understand where it comes from. This is Perm Stress' reply:

Originally Posted by Perm Stress

To daiquiri:

in this file I did modify Fr3 to have control edges intersecting hull and made an intersection. To point out, intersection line is made a knuckle, and a curve added. In order to have proper ends of this intersection line, certain Fr3 control edges would have to coincide with keel and deck lines.
Now you can delete parts of fr3 you do not need.
However, all this trickery will not go far beyond creating nice looking views in Freeship.
I for one, would not use this for anything requiring iterative process.
And would not use this kind of lines for design, except when say, bulkhead is not parallel to one of the four planes freeship can make intersection lines -frame, waterline, buttock, 45degree diagonal.
If I need to use this intersection for further work, I would make the control net on fr3 incredibly dense -due to the nature of surface description in freeship , any two lines will only be the same, when all their control points are the same, if this is a knuckle; if this is not a knuckle, surfaces containing said lines will have to be the same to the nearest knuckles. Otherwise we are in brute force situation -very dense control net on intersecting surface, loosing core benefit of delftship/freeship -few controls for any size of surface.

For possible results example, you can check "loyalist" -gaff schooner in Delftship site downloads section.

This is the file modified by him and refined again by me, in attempt to understand it better: Canoe-like boat mod.fbm

As far as I could see, the Layer Intersection function works in Freeship only partially... In a sense that it will find the intersection only if it is located between the inner control points of the two intersecting surfaces. It doesn't consider the outmost part of a surface (the perimetral band between the boundary and regular edges). That's why a coarse grid of control points is necessary close to the boundary edges, if one doesn't want to loose that part of the intersection line.

All in all, it's is clear that the Intersection function is not well developed in Freeship (and I guess in Delftship too, since they share the same surface-management engine) and that it is much better to proceed in a standard way:

- create the hull with regard to hydrostatics, stability and powering requirements with Freeship / Delftship
- export it as IGES file
- complete the job with Rhinoceros or some other true 3D modeling software.

That's of course, while waiting for VictorT or Martijn to develop a function for creating internal framing...

Cheers!

[frank smith]

daiquiri , which version of freeships does that file open in ?

[daiquiri]

I am using the latest - ver. 3.21

You can find a download link here: Free!Ship 2.6+ vs Prediction resistance by Holtrop-Mennen started method

[APP]

Originally Posted by daiquiri

Another question which I'd like to discuss is: how do you create concentrated masses with Freeship? Like the engine, tanks, deck items etc.
Cheers

Hi,

I am expecting to see that discussion

Regards
APP

[masrapido]

I might be wrong, some years have passed by, but Martijn had described his program as "hull modelling" software. Not as structural design program. To create frames you would need different software. That is why freeship has a choice of export formats.

Freeship was never intended for structural design, as I understood it from the beginning. Commercial version, which comes with additional extensions for a range of calculations, doesn't have the ability to create frames either, which confirms the above.

I just experimented with the latest attachment and noticed that no matter what mass I made any of the frames to have, the Cog did not move for a milimetre. With three frames astern, having combined 150kg, I would expect the Cog of the hull to move backwards. It didn't. So this approach may be futile if one is creating frames for mass calculations.

I create frames as parts and import them into the drawing. In fact, I create every mass as a part and place it in the position for CoG calculations. Cylinder is quite handy for that.

[yipster]

Early sunday morning in bed with caffee and my mobil so forgive type and other errors
Exporting to other programs has more advantages like easyr handling
For weight i think Martijn made a plugin check his site couse i.m not good in opening multi sites simutaniously on a mobil yet

[groper]

Ok daquiri, heres how i create frames in freeship.

First i open a new model with nothing in it. I add a new surface in the transverse plane at the zero point, make it wider and taller than the max beam and height of the ship i plan to construct. I add HEAPS of control points in both vertical and transverse directions with connecting internal edges so it forms a grid on the surface. The tighter the grid, the better the resolution you get when intersecting the edges (more on that shortly)
Save this as a "part" from the menu options, call it "bulkhead" or "frame" etc...

Start a new model and begin constructing the ship. Once completed and faired all the hull boundary surfaces (hull and deck etc) i then "import part" from the menu and import the frame template we created earlier. It will bring the part into a new layer (if it doesnt do it automatically, do it manually). Then select the entire imported part, and MOVE it to the longitudinal station where you want the frame or bulkhead by numerical longitudinal translation - which is why it was important to have the part zeroed before saving it as a part.

Once is longitudinal position is ok, select "intersect layers" from the menu option, and it will add new control points to the bulkhead where the hull and deck surfaces cut the interior edges of the "frame part" thus giving the transverse shape to the otherwise square grid. This is where it was important that the original part has a fine enough grid or you will get poor definition of the curved hull/decks surfaces with very few control points... Connect the new points via adding edges and turn the new edges into knuckle edges. Once the entire profile is defined by a knuckle edge all around, you can delete the rest of the part by selecting edges outside the knuckle edges and hitting delete. Now you have your trimmed bulkhead or frame and you should edit the density and thickness properties so that it will calculate the weights etc and you can use it in your CoG analysis and or modify it if needed later etc... Make sure you UNTICK use for hydrostatics in the layer properties, or these will screw up your hydrostatic calculation and you get the stupid "donk" error floating point BS garbage...

Repeat the above steps for all the transverse frames in your ship etc.

Adding concentrated masses....

simply create another part and import it into its own layer, like we did above - but its shape or size and number of control points is not important, only its location needs to be accurate to calculate the mass moment and distance for CoG analysis etc... For convenience, i try to make the surface exactly 1m^2 because i can figure the densities and thickness easier this way... ill explain shortly... Call the new layer "engine" or "whatever concentrated mass you like!". Now you can move the layer around by itself edit it etc, and most importantly, turn the layer visibility off so you wont see it ruining your otherwise nice looking ship...

Once in position, simply change the thickness and density of the new layer properties until the weight is what you require for the concentrated mass. It shows you the new mass realtime when you update the field, so this can be done quite quickly "on the fly" by trial and error. For small craft engines and the like, i usually end up with a density of around 10,000 or something like that and the weight will end up a couple hundred kgs etc... Often existing surfaces can be edited in this manner if the concentrated mass occurs central to the existing surface, simply add the extra mass to the existing layer mass and keep editing the density until you reach the summed total you were after.

Some things to consider;

Once youve created frames in this manner, they are not good enough to use in a real boat construction - i mean an expensive boat that is. The transverse frames will not have a good enough resolution to make a fair hull. So, I have another way of doing this but involves the use of more CAD software... A better way of obtaining cutting frames for mold building or actual boat frames, is to select "export" the "polylines" to 2D DXF file... it will export the polyline profile of each station taken directly off the NURB hull surfaces, not the frames... It will export the stations profile lines you see in bodyplan view... then open this DXF file in another CAD software to manipulate the frames, add offsets or whatever you need, then these can be sent to CNC router for accurate cutting etc... make sure to check the resolution of the exported polylines tho, sometimes it exports jagged lines which will also be no good for an expensive boat, theres an option to select resolution when exporting, im sure it works properly tho? I think it may be to do with the views "precision" setting... make sure it set to "highest" before exporting.

Hope that helps?

[daiquiri]

Thanks for the explanation, Groper. Looks like one possible method for the preliminary estimation of the hull weight.
Considering all the complications, I think I'll stick to the proven method - hull shape with Freeship, framing and other details with Rhino.
Cheers

[groper]

Yes, freeship can be very frustrating sometimes... its missing some really useful features and is very clumsy trying to do many things with it... its nice sometimes tho, because it can be very fast to model simple shapes for simple analysis... once the design moves forward to more detailed stages, its best to use different software!

[APP]

Originally Posted by groper

Ok daquiri, heres how i create frames in freeship.

... then open this DXF file in another CAD software to manipulate the frames, add offsets or whatever you need, then these can be sent to CNC router for accurate cutting etc... make sure to check the resolution of the exported polylines tho, sometimes it exports jagged lines which will also be no good for an expensive boat,...

Very nice tutorial. thanks. By the way when you have a layer of the hull, we need to subdivide it for the dxf/cnc to create the single shell panels/plates to weld. What are usually the dimensions of each panel/plate if it is developable? Is it in function of the frame span? How do you determine anyway the max size of the panels (I think are also called strakes). I hope my question is clear.

Regards
APP

[LP]

And another method:

I extrude a short panel (0.1' or similar in metric) from the centerline/keel edge. I extrude again to deckline or there about. Then, I do a transverse extrude to desired width from the previous fore and aft edges (multiple bulkheads). The first two panels can be deleted leaving your roughed in bulkheads. At this point, I will add an intersection curve to the desired station if it doesn't aready exist.

On a chined hull with a control edge and points coincident to the station, I will add matching control points to the rough bulkhead, tick the "corner" box and lay them over the hull points. (You may know how to do this, but I'll throw it in. SELECT the point to be moved, Hold the CONTROL key and SELECT the destination controil point. TAB though the X, Y, and Z boxes and your points will be co-located.)

On a rounded hull with control points on the hull at the desired station, I'll repeat the above procedure for the sheer edge and keel line. At this point, I turn off all layers except for the bulkhead/frame and leave the intersection curves turned ON. Then, I manually insert points and move them so the bulkhead edge coincides with the intersection curve. It's not computer automation, but it gets the job done.

For contentrated masses, I simply draw a 1' square centered at the proper location, give it a thickness of 6" (sym = times 2 = 12") and make it's density equal to the desired weight of the mass. I'll leave the metric conversion to you.