We have been using glued contacts to join together sturctures of shell elements. In particular to represent lines of rivets. We have found the results to be unreliable, both statically and dynamically. Has anyone used these contacts with shells in models with verifiable answers? Carlton.
"Carlton Matthew" wrote: > >We have been using glued contacts to join together sturctures of shell elements. >In particular to represent lines of rivets. We have found the results to be >unreliable, >both statically and dynamically. >Has anyone used these contacts with shells in models with verifiable answers? >Carlton. Carlton, I'll repost the answer I gave Wednesday to your last question:
>---------------- Carlton, This can be an issue using the original glue formulation (penalty stiffness) if the glue surfaces are not coincident. This type of glue is similar to using a CELAS spring between non-coincident nodes. NX Nastran 6.1 has a new glue formulation (which is the default) that uses a connection method more like a cweld, so it can transfer moment between non-coincident glue surfaces. using this new method, you should get 6 clean rigid body modes on a glued free-free structure. Regards, Jim >----------------- Note that when shells are modeled at their respective centerlines, you will have the case where the glue faces are separated by a distance equal to the sum of half the thickness of each shell. Please respond with an indication of which release and which gluetype you are using. Also note that if you are trying to model discrete rivets (rather than a continuous bond), it may be more appropriate to use CFAST/CWELD elements instead of surface gluing. Regards, Jim