Greetings,

Short version:
I am somehow messing up sheet/shell surface-to-surface contact (not glued)

Long version:
I am currently using NX Nastran 7.0 and dealing with shell contacts.  I made a simple model of 2 shell sheets, .5" shell thickness value, stacked on top of eachother (space between sheets is .5").  I did a simple compression run (sheets in the XY plane, vertical 'up' is +Z) with a small force in the -Z direction.  The only way I can get it to run is with glued contacts.
The error I get in the F06 file when using non-glued contact is:
USER   FATAL   MESSAGE 9143 (SUPER3)
THE SOLUTION FOR THE RESIDUAL STRUCTURE DOES NOT EXIST.

The exact model with glued contact runs fine and gives good results.  The only parameter I am setting manually is AUTOSPC=1, everything else is straight default.

I have tried constraining the top shell sheet in to travel only in the Z direction, but this did not help.

A run with 1 sheet and 1 solid works fine with surface-to-surface, non-glue contact.

I knew I was in trouble when I didnt find anything here about the subject (longtime lurker).

Any help on the method or settings I am goofing up would be appreciated.

Thanks,
Dan

Are the individual parts constrained in 3D space to prevent rigid body motion? You cannot rely on the contact surface to provide the rigid body motion.

You could try a static and/or buckling (eigen value) solution determine if you have rigid body motion.

Brian
www.espcomposites.com

Did you defined the contact region surfaces manually or automatically?
Did you defined the Global Contact Parameters in the BCTPARM / Case Control Tab? If not you have to define those parameters as the number of iterations, initial penetration etc... You might also try to reverse contact directions bu changing between the source and the target. NX is surprisingly good when it comes to NOT WORKING and output-ing ERROR messages.

There is a little movie in the NX HELP where it is shown an example of making a simple contact, surface to surface between a shaft and a wheel. You can look for it and do the exact same steps and the same settings...It worked for me..sometimes...other times i did the exact same steps and i get a error message. I delete every thing and start from scratch..and it works...

Hope this helps

ESPcomposites and Bailos - Thank you for the suggestions.  They led me to find the SILLY mistake I was making.  I neglected the offset between the shell surfaces (manual contact, Edit Regions menu) (I left the offset zero in the shell definition our of ignorance for the process).  The gap betwee the shell surfaces must be accounted for either in the shell element definition or the contact definition.  Duh.

Thanks to all that read and all that responded.  I apologize for the long question from a silly mistake.

Well now I feel like a darn fool.  It wasnt the offset, it was the proper labeling of the TOP and BOTTOM surfaces in manual contact settings that did it.  I evidently didnt do that the first few times, either.  D'OH!  Now I really look silly.

I still cant get it to cooperate for the automatic pairing contacts...

Thanks again everyone!

Dear Dan,
Things to check when defining maually "surface-to-surface" contacts with SHELLS:
• Contact distance definition in the contact property should entered correctly: in this case use 0.5 or bigger, say 0.51 is correct.
• Surface orientation: Shells elements have TOP/BOT faces, then SOURCE & TARGET regions should point at each other, if not contact will not be success.
• By default, shell offset is on, then NX NASTRAN accounst for Shell thickness in the contact interations.
• INIPENE parameter accounts for initial penetration: this is the most important parameter to define correctly in surface-to-surface contact, by default is defined as "CALCULATED", this means that if NX NASTRAN solver find any penetration between source/target contact elements then will resolve the "interference-fit" situation, and you will see locally high stress, then mind your steps, OK?.

With GLUE surface-to-surface contact, the orientation of shell elements is not critical.

Best regards,
Blas.

~~~~~~~~~~~~~~~~~~~~~~
Blas Molero Hidalgo
Ingeniero Industrial
Director
 
IBERISA
48011 BILBAO (SPAIN)
WEB: http://www.iberisa.com
Blog de FEMAP & NX Nastran: http://iberisa.wordpress.com/

BlasMolero-

Great tips.  Thank you very much!

Dan