Nastran/Patran Contact Analysis

[40818]

Hi,
I have a honeycomb sandwich barrier (aluminum skins and core) which is held in say 4 positions. I wish to apply a pressure load to the structure and get some decent results out. Now on the opposite face to the loaded side their is a flat structure which my barrier is allowed to contact, and when it deflects and cotnacts the surface it will give different load distributions to the normal 4 pick-up points. Now say the gap is 0.1 inch, i dont want to just SPC any point that deflects above this limit, as i dont believe is is the right way to go about it, which leads me to a contact analysis where the nodes are allowed to move a certain distance and upon reachin gthat threshold they are effectivly held. I would require to be able to then get out contact forces as well as the 4 pick-up point forces.

Any ideas on the best way forwards?
Thanks.

 

[cbrn]

Hi,
I'd suggest non-linear springs with almost-0 / almost-infinite stiffness with the "knee" at the deflection value corresponding to the gap you have...
In ANSYS there would be at least three or four different ways to do something like that, but in Nastran I don't know.

Regards

 

[cbrn]

Hi,
sorry, my post was incomplete: the reaction forces of the contact would be retrieved on the auxiliary "totally-restrained" nodes at the end of the non-lin "springs": nodes showing almost-0 reaction force mean "no contact".

Regards

 

[40818]

I would ideally like it to be a linear analysis, and have found out about using "linear Gaps" or "Spoints" which it seems will do the trick.
Now i just need to try to work out how to use them correctly!

 

[rb1957]

contact problems are, by their nature, non-linear.

are you worried that the contact will overload the second structure ?

is the bulkhead able to react the loading based on the 4 attachments alone ?

i'd consider the problem from a different direction, is there anti-scuff protection on the potneital contact area, 0.1" isn't much space ?

 

[40818]

It seems the MacNeal Swindler Corporation have a method whereby you can use a linear gap element so you dont need to go to a non-linear analysis for contact, so its good for simple contact analysis such as heel toe effects of direct compression of faces etc. I have a powerpoint pres of theirs which gives the quick theory and usage, but have to try to work out the "how to do it" now.

I'm not worried about the secondary structure being overloaded, but i will need to go through the motions as it were. The 4 attachments are able to carry the whole load, but as its applied as a pressure i get deflections which will introduce the contact. I'm not worried about the contact in itself, it is designed to do so in that loadcase, as its a 9g crash load. Conservativly i could size the 4 attachments to carry all the load no problems, its more of the contact area/force i would like to know.

 

[cbrn]

Hi,
OK, the algorythm may linearize some parts of the problem, for example by not taking into consideration the contact stiffness update and so on (which, for a "remote control" or a "gap element", makes sense). But, as rb1957 points out, a behaviour such as the one you describe, "allowed to move a certain distance and upon reachin gthat threshold they are effectivly held", IS non-linear by definition. Any full-linear resolution of something like that will be dangerously UNREALISTIC (at best...). So take care...

Regards