How to create a rigid picture frame (square) with rotating corners ?

[RIST]

I am trying to simulate a picture frame test, in order to study wrinkling in thin shells. I found a thesis where its done something similar, but i am not able to understand how to actually do it, since i keep getting errors in element section with rigid wires ( that cant be assigned sections).


Would somebody explain to me how to emulate that rigid frame, i am new to abaqus and have not been able to understand how to make the parts, connections and constrains necessary to build a rigid frame as presented in this thesis.

Thank you.

 

[Mustaine3]

In the rigid body options is a possibility to tie reference nodes onto the rigid body. Use this to add a ref node at each end (rotation point) of each bar. Between those node you put connector elements.

But be careful to not create an overconstraint.
I would say in 3D you need 3 hinge connector and one connector, that locks the remaining two translational DOFs. That could be a Cartesian with rigid elasticity in those 2 DOFs.

 

[Mustaine3]

PS: 3 hinges and 1 slot should work

 

[RIST]

Thank you for your response. I was able to assign the connector elements as you mentioned, but still the structure doesnt move after applied BC, not sure what I'm doing wrong, probably i am doing something wrong in the creation of the connectors/wires/ref.points. The objective is to make a shell obbey to this movement, which i would then do with coupling, but if you have any sugestion, to make the edges of a plane shell behave as rigid it would also be helpful

 

[Mustaine3]

I can't tell you whats wrong in your model. Maybe the little example I've just created helps.

 

[RIST]

Thank you so much for the help, you example is definitely what i was trying to do, i am now using your model in order to learn how to do it my self, since I am having some issues in coupling ( probably because your frame is only mesh a not geometry) with my shell, but, this is already very helpful. Again, thanks for your help and time lost

 

[RIST]

Something is still wrong with my modelling, even though I try and copy everything you did, could you take a look at my model ?

 

[Mustaine3]

I have currently not enough time to debug your model, but I've attached my .cae file, so you can compare the definitions by yourself.


Little trick: Translate some instances a little bit in Assembly module to ease the creation of the constrains and connectors. Afterwards you can move the instances back to their original position. The definitions don't change with that. See my other model in the .cae file as an example.

But there is one limitation: When doing that, you just have to prevent using the option, that creates local orientations in the connection builder between the two points. This method takes the locations of the two points. When you move the instances back and the two points have the same coordinates, the update of this method fails. The workaround is, to reference on a existing local CSYS in the connection builder. Then it is no problem when the two nodes of a connector have the same location.