Hull fairing software (big ship)

There were many discusions about fairing software but almost all projects on many software sites are yahts boats or simple ships. I have a problem because I have to repair and fair hull surface of large container ship with complicated fore section with curved bulb. Have you got any experience with this kind of fairing. My problem is also difficult because I got this surface in IGS file from ship model basin and it containes 1300 meshes which are not perfect in few moments (i found an article on Rhino site about problems wiht IGS files and fairing but they all are ineffective for big complicated hull (or mayby I'm looking for too easy solution )

I was trying to create one healthy new surface but Rhino is always complaining near bulb frames...

 

You can fair a surface in rhino for the hull and other surface for the bulb and then try to connect them with a network surface, you can either use push and pull technique, or loft trough section cut form the model, I will try first to approximate a surface to what you have and model the bulb apart, using just push and pull to fair you surface.

other maybe quick solution can be use prefit from maxsurf maybe having some section of the hull into this program may help to recreate a nice surface, the problem here is that you have an iges polysurface that hardly will be able to be used properly in most programs, I hope any user form maxsurf can comment on this one, if is not a problem I can have a look at your file if you like just to see if I can approximate such surface in rhino.

 

Hi,
I have some experience in fairing hull curves for large ships, but I use Tribon M2 SP4 Initial Design for this. In my opinion try to find some profesional software for large ships, or ask for offset table, because how to look this surface is not right and not accuracy. The pictures from attach files are from one container carrier that I work , but I finished the surface and in this moment I work at shell plates and profiles and expansion shell plates. I think the CGN solutions are reasonably. In my work I meet this situation and some time I split the surface in 2, 3.., sub-surfaces, connected each other, some time I create the section, buttock or waterline curves intersecting the bad surfaces with planes, I faried intersection curves(sections/buttocks/waterlines) and I recreate the paches. It is important to choose what intersection curves to use for fair and after this to recreate other curves around the fair curves.

 

Thank you for your comments.
I know the method of fairing hull but first I have to solve multi mesh surface from IGS problem.
I have Rhino and SC2004 with ShipCAM (Loft). Unfortunately to create one surface from few selected ShipCAM ask me for "Yard" licence - also I don't know if the result of this function is good for me.
I was trying many methods in Rhino to create surface but they all are failed near bulb section. I was cutting IGEs meshes by frames waterlines etc trying to connect it by loft or network surface function - results were ugly (I don't know where Rhino has brake line function similar to ShipCAM.
We don't have MaxSurf (from CGN suggestion) to try his method.
Do you have any suggestion, which can help me - I waste my time going around since Monday?

 

Maxsurf has no problems doing bulbs such as the one in your example (see attachment), however your major problem is the IGES file with the large number of patches. Unfortunately a lot of hulls done in general purpose CAD systems suffer from too many patches, too many control points, poor fairness and a large number of gaps between surfaces. Even hulls from NAPA and TRIBON tend to have this problem, albeit to a lesser extent.

There is no simple way to convert the IGES file you have to a smaller number of surfaces and control points, you may as well create a dense table of offsets from the surfaces you have and start creating a new hull from scratch.

Prefit is not suitable in this case as it more suited to simple fair surfaces and will not handle bulbous bows. We do have a genetic algorithm based fitting function within Maxsurf that may be useful, if you send me an IGES file of NURBS surfaces I would be happy to try it for you.

 

In fact my IGEs file was generated in NAPA. We are doing our job in ShipConstructor but with this file work is somewhere between difficult and impossible

I have to add few technological details to this surface and then fair in that places where fairing will be needed - now it is quite good.

For ShipConstructor (and other soft) there are to many surfaces which generate curves consists many simple lines which begins and ends on small surfaces edges, also they have invalid 'sides' or 'ways' and correcting this manually will take few weeks.

Andrew where I can find your mails address ?

 

If you wish to send me your IGES file I can be contacted at AndyM at formsys dot com

NAPA constructs it's hulls from a large number of bicubic patches (effectively 4 x 4 control points) and as a result taking those surfaces into other programs can be difficut. It is better suited to lines fairing system (i.e. fairing existing hand drawn lines) than the ab initio creation of hull shapes.

Maxsurf on the other hand was intended to create designs from scratch, and as a result is structured in such a way as to use the minimum number of surface patches and control points. Of the designs shown in my previous posts all can be modelled in Maxsurf with between 5 and 10 surfaces.

We have had a lot of cases where we have been sent IGES files from other CAD systems that we have recreated in Maxsurf. In one case, a frigate design done originally in CATIA, we were able to recreate the design using about 1/100 the number of control points. The Maxsurf hull was fair, while the original was not, and had no gaps between the surfaces, which was not true of the original.

 

Andrew,

If you remove 99% of the data points that define a surface it will obviously be easier to fair but how well will the new surface reflect the original? In other words, how to you decide which points aren't important and how do you determine the quality of your reproduction.

I've been wrestling with this dilemna in manufacturing for 20 years and I think one needs to have an awareness of the end function of the surface created and the knowledge and ability to select the appropriate point on the continuum between perfect reproduction of the original, usually inadequately resolved, datapoints, that often define an unusable surface and retaining the 3 points that are the minimum required to define the infinitely fair surface.

I have no experience with your software so I am making no value judgement or comparison.

What I find useful in Pro/E is the ability to include all original datapoints, create lines (splines) through the points selecting a smoothness/adherence to points bias, using those lines (curves) to define surfaces and again choose a smoothness/adherence to curves bias. At that point I can go back and review the original points against the derived surfaces and often recognize points that represent either incorrect data entry on my part or simply violate the fairly, no pun intended, obvious original design intent and adjust or remove them as appropriate.

All of this may be more important when bringing analog aerospace data into the digital world but I also believe it relevant to these discussions.

Doug Carlson

 

Doug

In the case I mentioned we were not removing 99% of the data points, we were doing a least squares fit to the existing data using a simpler surface. Obviously you have to make a value judgement as to how many control points to use, however this was fairly easy as we could see that the hull was actually very simple and was supposed to be extremely fair.

The Genetic Algorithm based fitting code we have is quite different to the standard skinning algorithm that is used in most CAD systems, in that, in addition to trying to minimise RMS error between the surface and the data points, it has control point net fairness, spacing and orthogonality as additional objectives. This results in a net of control points that is very similar in layout to what a human designer would choose. This is very different from most surface fitting programs which produce awful control point meshes that are totally unsuitable for further modification by a human designer.

I've attached an example of the GA fitting output, you can see that the control point net is very regular and is spaced the way a good designer would have done.

 

OK new week and new ideas
I found some rule in ShipCAM which can probably solve my problem. Maybe it is not best solution but it is only thing I have now.
In ShipCAM when I imported IGS file I've got many surfaces. But when I was saving this in ShipCAM format it generates many scratches in all directions at short distance – it was useless.
I found the rule that ShipCAM generate base lines correctly when meshes from IGS files are in correct order in correct way. I corrected this using transpose mesh and change direction functions. Then when I saved to ShipCAM all created lines were correct! Now I have to delete some lines because there are too many of them and generate new surface.
My problem is that I don’t have Yard license of ShipCAM to use Generate Surface function

 

I use SeaSolution for the lines fairing and I do not see any problems with bulbs etc... When you work with NURBS surfaces, you must decide in beginning how to split your hull into surfaces, like fore body, aft body, flat side, and flat bottom. Moreover, this decision is very important for future fairing process, quantity of control points and time consuming. Very often, I receive preliminary lines from different systems and my job is to make final production lines fairing. I never use IGES file with surfaces from NAPA. I prefer to use 3D lines. Its take little bit more time for lines fairing, but in the end I have flexible model. In attachment, you can find some ideas ho to make NURBS surfaces fairing better and acceptable for the most different systems.

 

I think that I can finish my first lesson with this hull. My solution is to ‘merge’ some meshes in Rhino (FOB, FOS) and then import and correct some of them in ShipCAM to work with structure.
But I think that it was too difficult to my first lesson and I have to choose easier hull shape at the beginning (boys put me in to the deep water - next time I have to ask my friend about another old hull surface and I will start with offsets and scratches not IGS multisurface).
And here I have a new question. Is it possible to move smoothly from fairing for example waterlines to frames, from frames when working in Rhino? How to create correctly surface when ‘Loft’ function is going crazy near the bulb and ‘Curve Network’ don't want to create surface which I want to section in next direction - maybe there is another way ? (of course I have read PDF from Rhinos pages about fairing but there is very simple hull shape).

Thank you Alexander for your document.

 

Ship surface shape is quite special and some times with very special requairements. That why software with standard set of surface manipulation tools can't help so much. In the middle of may I have plans to visit Gdansk with SeaSolution software demonstration. Welcome. You know my E-Mail.