S.O.S to FEA experts

hey all...need to do an FE model for an OSV with all structural members...I have all the required 2D structural drawings...Initial plan was to model it in ShipConstructor and to see if its possible to import the geometry into ANSYS to do the mesh and link ANSYS with solver LS-DYNA...

well always murphy's law prevailed and the shipconstructor network licence not working...now thinking of doin the entire structural modelling in ANSYS..has anyone done a complete structural model in ANSYS ??..

is it too much of a pain in the backside ?

any heads up ?? alternatives ??

i have access to ANSYS, LS-DYNA, solidworks, PATRAN, autocad, maxsurf workshop..

btw need to model everything as shell elements of its gonna be ages for my supercompuer to finish the simulation !!

all help appreciated..

 

Depends upon your patience, your time frame for completing the job and how accurate you want the results.

I do all my modelling inside the FE program. Takes a bit of time, but I control the input and it makes for manual meshing much easier.

If the program is not user friendly enough for you to model with...choose another.

 

thanks a lot ad hoc..i think i will do the same...i have time...which FE software do you normally use ??..i plan to use ANSYS...

 

We use Cosmos since we do our design work with Solidworks, which I noted was on your "list". We've been very pleased with that program, having used Algor for many years prior and having all the usual meshing problems that are generally absent the Cosmos package..

 

Ditto.

Ive been using Cosmos since 1993. It is easy enough to model inside the program, as it has a basic CAD generator within the program. Something many FE programs didn't do until the early 00's. Cosmos , for me anyway, is very intuitive...unlike many.

 

I'd like to plug a package called MAESTRO.

http://www.maestromarine.com/

It's already set up to handle ship-type structures, specifically stiffened plating and the like.
With general purpose FEA packages, modelling vessels can get interesting, especially when it comes to modelling loads (hydrostatic support of the hull, fluids in tanks, etc.)

 

I'm curious: why do you want to analyze the entire vessel with FEA? Is their some novel feature in this design for an OSV which makes it radically different from other contemporary designs?

It helps to know what you expect to get out of the analysis before going to the trouble of building the model and generating the loads.

 

@NavalSArtichoke : I need to actually model the OSV till the fore collision bulkhead and also the mid section of the vessel [for bow impact and side impact ]...the rest will be treated as rigid and structural details will not be modelled, just the hull will be treated as rigid....

the objective is to use the FE model for OSV-offshore jacket collision analysis using LS-DYNA.. at the moment, u would probably know tat there is a lot of interest in this subject from the point of view of offshore jacket damage and also the vessel damage from the point of view of damage to fuel tank and environment damage thereof..

not just OSV , need to model few other types of vessels after risk analysis, to understand the consequences...

its part of a wider industry supported project in europe...as its difficult to get a hold off of structural models, so building them myself...tedious job I understand after just a few days of it !!

any tips appreciated...

 

I see. Until recently, most OSVs have been designed to classification rules and as far as I know, not a lot of FEA has been done, which would explain the lack of available models.

In the Gulf of Mexico, as platforms move further offshore and have become so expensive, OSVs are not allowed to moor directly to the platform as they did in olden days. More and more operators are demanding that the boats servicing the platforms be equipped with Dynamic Positioning Systems so they don't have to tie off. Some older boats have had DPS retro-fitted, but a large number of newbuilds are having it installed during construction.

 

Hi - Most of my preprocessing is done in rhino then I import into Strand7 . LS-Dyna I have used many years ago for a metal forming project. I can strongly recomend using rhino for organising and subdividing geometry or building geometry for use in FE. Just set the tolerance to a suitbaely high setting as the default tolerances can get you into trouble. If you use a parametric modeller say solidworks then you can export the geometry to rhino (via step) then to LS-Dyna. The surfacing features are usually a little rough in parametric modellers, but some are good eg catia. Hope that helps. Peter