Zwicky
Marine/Ocean
- Jan 28, 2013
- 7
Hello everyone,
I'm currently working on a 40 meters ship model for linear static analysis, mainly all shell elements QUAD4 (small thickness structures). The geometry has been prepared with rhino in order to use the new Femap V11 improved edge-surface glue feature.
First of all I made a couple of tries with femap v10 to see if it was really not working properly, as said by siemens' salesmen, and it actually seems right, i noticed some relative displacement where it shouldn't have been.
My question is this: my model has approx. 400 connectors, and i know that nastran is not very happy when same elements belonging to a surface are attached on both sides (this by the way for a deck-like surface is not very nice,
because it is laying on the hull frames and has deckhouse or other structures on top of it, so this sounds like an omission from salesman in my opinion), but going on with the modeling, some connections that were working perfectly just start to fail,
without me modifying any of the associated elements, properties, entities, etc. etc.. Does someone have an idea why that happens?
To be more precise i use glue-contact connection type, glue type ''weld'', glue factor ''1'', search distance of 3.5 mm because all the shell haven't more that 6 mm thickness, and i repeat there
are only edge (curves) on surfaces. The sources regions are surfaces and the targets are curves (also this was adviced by femap training partner but i read on nastran help that actually the source should be the curves and present also a finer mesh
because nastran creates edge elements to model the connection). If it was not clear until now, i have a big chaos in my mind about this topic, if someone had more experience about this and can give a clearer explanations please give your advice,
I would be glad to hear from you.
Best regards,
Zwicky.
I'm currently working on a 40 meters ship model for linear static analysis, mainly all shell elements QUAD4 (small thickness structures). The geometry has been prepared with rhino in order to use the new Femap V11 improved edge-surface glue feature.
First of all I made a couple of tries with femap v10 to see if it was really not working properly, as said by siemens' salesmen, and it actually seems right, i noticed some relative displacement where it shouldn't have been.
My question is this: my model has approx. 400 connectors, and i know that nastran is not very happy when same elements belonging to a surface are attached on both sides (this by the way for a deck-like surface is not very nice,
because it is laying on the hull frames and has deckhouse or other structures on top of it, so this sounds like an omission from salesman in my opinion), but going on with the modeling, some connections that were working perfectly just start to fail,
without me modifying any of the associated elements, properties, entities, etc. etc.. Does someone have an idea why that happens?
To be more precise i use glue-contact connection type, glue type ''weld'', glue factor ''1'', search distance of 3.5 mm because all the shell haven't more that 6 mm thickness, and i repeat there
are only edge (curves) on surfaces. The sources regions are surfaces and the targets are curves (also this was adviced by femap training partner but i read on nastran help that actually the source should be the curves and present also a finer mesh
because nastran creates edge elements to model the connection). If it was not clear until now, i have a big chaos in my mind about this topic, if someone had more experience about this and can give a clearer explanations please give your advice,
I would be glad to hear from you.
Best regards,
Zwicky.
