AutoCAD to ANSYS

[issac82]

Hi,
I am new to this major.I want to ask that I have an AutoCAD drawing of a ship.I wanto to make a ship model using this drawing in ANSYS.Can somebody tell me how can i do it?is there any software code which can also do this?thanx for ur advice in advance

[CGN]

Use the Autocad file to create a 3D model in any 3D Modeler like rhino or a hull surface modeler like Maxsurf then export your file to Ansys as an Iges, Parasolid or ACIS file.

It may not be easy for being the first time, but It will be better to have the 3D model of the hull and then import into Ansys

If you need help on how to model the hull in 3D search the forum there are good references available, also Rhino website has a good tutorial on how to model a hull.

[issac82]

Sir with due regard plz can u help me for paste some links here for some helpful material regarding this.I have gone through some references but they are too diverse and still i am unable to find the exact material for this purpose.

[CGN]

Basic tutorial on how to loft a hull:

http://www.3drender.com/rhino/boathtml/index.htm

This is for solidworks but the principle is the same for rhino (lofting is the same for any software) just pay attention on how is made and use rhino to follow the technique

http://www.youtube.com/watch?v=GS31uJczzvo

This is for Turbocad but is lofting so the same technique or similar is used

http://www.worldwideflood.com/CAD/hull_loft_tute.pdf

Also this is a good tutorial
http://www.rhino3d.com/tutorials/fairing.pdf

http://en.wiki.mcneel.com/default.aspx/McNeel/RhinoTutorialLinks.html

Once you have your hull you can import into Ansys and for that the manual should be ok to guide you on how to import the model

[issac82]

thanx a lot for this valuable advice.I have come to now that somebody also used the AutoCAD ship lines plane and made a model using Gambit and told me that it is very easy in that also.do u hav any idea or somebody can guide me about that also??

[CGN]

No doubt the best way is always to model using the actual FEA or analysis software but for a hull not sure if all the tools are available or "friendly" to use to model the hull

[kilbysg]

I saw a movie clip once on the youtube. But they were using solidworks.

[mesut]

hello everbody

i have a 3d ship hull with bullkeads and i wanna import it into ansys but i have been trying to find the appropriate format in order to avoid problems coming from ansys but i could not.i have tried İGES but somehow it tends to bring problem as importing into Ansys.Could you help me? what is the appropriate format for ansys .Thanks in advance.

[yipster]

i'm trying inventor, ansys fea build in
there are more fine ansys modules than fea; look under products http://www.ansys.com/products/default.asp
MikeJohns pointed me at http://www.strand7.com/ looks good but does not import nor has library's
just in my e-mail http://www.comsol.com/ with free CD http://www.comsol.com/intro/structural
but now i'll swith this apparatus off eat something and watch tv

[Ad Hoc]

You have to be careful when importing "lines" from one software into any FEA. Even if the 'translation' works using either IGES or DXF.

You can have either one point or literally millions of points per curve. This just means you have to do a lot of "editing" of the data inside the FE program. Because to create the FE model requires connectivity, ie lines that all join together and the intersections being real intersections to model structure.

I gave up importing lines from other software years ago and now just model inside the FE program. Much quicker and easier and i can control where the line breaks are for ease of creating surfaces or extruding volumes etc. Since without this, the mesh is a mess!!!

[mesut]

Hello all,

Thanks to you all for your help.İ was thinking of modeling it in Fea programme.
i guess i have got to extract the relevant keypoints from 3-d model modeled in rhino then put them in Fea to model lines and areas, isn't it? to model the entire hull is gonna be tough:)

or if you have better ideas to tell me ,i appropriate that.

Kind Regards,
mesut.

[yipster]

calculating a single element is called a problem, and can be done by hand (where is that ME in a day handbook?)
non line forms can only approach the real mechanical behavior and multiple elements cant be done by hand they are called continues problems

inventor ansys uses as little lements as possible but can use over 30.000 nodes but just as pleasant to test as strand7
when using convergence adaptif meshen means having the mesh smaller at stresspoints but makes calculations longer

material strenght library's can also be used on more complicated meshes but the fea build in is a basic one
the book i'm using mentions at least a few months study for each ansys module, so i'm learning and now also order that CD

[Ad Hoc]

FEA is not simply...oh look, software with nice graphics and i get numbers out too...one requires a significant understanding of structural design to use FEA properly, otherwise you'll make many many mistakes.

[alidesigner]

I used to frequently model 50m cats and monos using nastran with Femap preprocessor, now called Femap with Nastran by UGS.

I took the hull surface into it via iges (bottom, chine & sides only), used the split command to subdivide the surfaces at frame sections then built everything else in Femap as surfaces. I then just automeshed the surfaces. Took about 2 weeks from start to finish.

Due to the simplifications you can make in fea, more often than not cad data is not suitable for fea modelling.

Forget strand, you need a geometry based modeller with its own set of drawing/modelling tools so you can make quick associative changes to your mesh. Strand relies too heavily on external imports (or it did when I last looked)

FEA that's bundled with 3D cad like cosmos in solidworks and ansys in Inventor are usually optimised for solid models whereas for boats you need shell and beam element models (havent looked at Inventor bundle but solidworks one was no good for boats). I hear that ProE and ansys work well together though.

Ansys (stand alone) and Femap/nastran are the only two I would touch, and of those two feamap/nastran is much cheaper and easier to use.

[Ad Hoc]

I use cosmos, and only model in cosmos. Plate and beam elements are all that are required, for basic FE model. However, if you require 3D models for 3D elements, then you really need to know what you're doing and understand about mesh sizes and their consistency. Regardless which input method is used!

[yipster]

Thanks for your views and advice Ad Hoc and alidesigner

sure, i'm painfully aware of not having a thorough ME background yet already in dos days i produced 3d cad drawings and animations that went on CD's packed with ME study books
but ok i did the drawings not made the lessons so there my cry where is that ME book

early nineties 3d pc drawing was a bit of a study but being brought up with drawing board and ruler pen it was great
FEA and CFD, another that fascinates me even more, were unheard of than
they are amazing new tools worth a study as if it were only to know what is and isn't possible

i made wind, explosions and smoke in 3dstudio but so far always had to make the paths or deflections and can imagine how by manipulating variables CFD does that even better for you
and how nice a tuft grid in a 3d drawing could be, all this is still limited i guess as i still have to see lift, drag, wind and waves, eddy's and what not visualized and calculated

at the marina yesterday i thought of your words of warning Ad Hoc as í'm not aiming for a certificate and, as with so many things, only experience gets one comfortable with a surrounding,
and to be honest, they are to be solved but i don't particularly "like" problems so i wondered what i'm actually trying to accomplish as i was more interested in taking a roller furler down

anyway, inventor works nicer as acad, back to the book

[Ad Hoc]

yipster

the way to think of FEA (and indeed CFD) is this:
Would you crack a walnut with a hammer or a nuclear war head?...if you choose the former, the good old hand calc's and basic education/training will get the job done and usually, not that difficult, and you get the walnut to eat. If you choose the later, then, as with a nuke, you'll make a huge mess with lots of 'fall out', walnut everywhere, not "usable" ...looks great though (as do the pretty colour plots)!...even though ones focus is actually cracking open a walnut...not the 'amazing looking fall out'.!!!...everyone is looking at the mushroom cloud, not the real results!

[yipster]

i hear you and forgive the explosions and smoke, that was from my "games" day's :o back to read

[Ad Hoc]

any questions...just ask....FEA is a minefield and not to be beguiled by the pretty pictures as many are when using FEA and CFD!!