Femap buckling question

[merida_2015]

Hi!

I am having a problem with buckling analysis.

I have a box beam made of plane elements clamped on both sides, but i need to apply force at one end in the middle of the cross section which means that i have to use a spider node. but, i am not sure how should i define element at the spider node...is it RBE2 or RBE3? i am doing something wrong because errors are showing up or i am getting wrong results. everything was ok when the beam end with force applied was free (not clamped).

Thank you in advance!

Sandra

 

[jbrackin]

The choice of RBE2 or RBE3 is important and should be based on desired behavior. RBE2 enforces "no relative motion" between the included nodes and therefore can actually make a section stiffer. The RBE3 enforces a "weighted average motion" and does not change the local stiffness of the included nodes.

If you are getting an error when you add constraints to nodes that are part of the RBE2 or RBE3, it is likely due to dependent DOF conflict and can avoided by setting the param "AUTOMPC" on the bulk data input form in Femap.

Regards,

Joe

 

[rb1957]

not following something sandra. the FEM was working before when the end where the load is applied was free, but now it's clamped.

how does this relate to the RBE question ?

another day in paradise, or is paradise one day closer ?

 

[FSB1]

From the information you provide it looks like you have a beam of hollow rectangular section - rectangular tube, otherwise I don't see the need for an RBE.

If you are applying the force at "spider node" then you should use RBE2 because like this
the load applied is going to be transferred from the master node to the dependent nodes. So this is the same as not having the spider element and
applying the load direction in the dependent nodes.
Also remember that using the RBE2 you cannot constrain the dependent nodes because they already have
a "constraint" imposed by the independent node. Otherwise a similar error will show up: "GRID POINT xx COMPONENT xx ILLEGALLY DEFINED IN SETS UM"

The use of RBE3 in this situation is not really correct because your "spider node" will be the DEPENDENT node. So you'd be pretty much defining your model "telling" it that the independent nodes depent on a node that depends on them (confusing!!)... your load wouldn't be correctly transmitted and that's why you get the wrong results even thought no fatal error show up because you are constraining independent nodes only.

My recommendation is: if you really have to use the spider, and have this end constrained, then:
-use RBE2
-constrain the "spider node" - this will constrain your free end
-apply the force in the "spider node"

If you want, compare the results with:
- no spider
- fix the end nodes
- apply the force individually at each node of the beam end

Results should be the same

 

[merida_2015]

Thank you all for replies, you really made my day!

I was reading a lot about rbe2 and rbe3, but it was hard to understand it percisely.

Leo16, i tried you recommendation and it went really good (i have some results to benchmark). i think that i get it now.

Thank you for your help!

Sandra