CBUSH elements with non-coincident nodes

[MBD26]

I have two parts (a skin and an aircraft former flange) overlapping with a row of fasteners along the length of the overlap modeled with CBUSH elements. The parts are modeled with CQUADs, and are slightly offset, such that the CBUSH elements have non-coincident nodes (though the CBUSH elements are in line with the axial direction of the fasteners). If I'm only trying to get fastener bearing and bypass loads from this model, is it necessarily a concern that the CBUSH elements have length? How would it change the result if the parts were flush and the nodes were coincident?

 

[pbd999]

You will have to decide if it would change the results in the fastener loads and if that change is meaningful, but I thought it might help to give my thoughts on this topic, which I find is a question that comes up pretty often where I work. It might be helpful to compare CBUSH to CELAS elements with non coincident nodes. A CELAS spring simply connects two DOF with the K specified. It doesn't understand geometry at all, it just connects the nodes you specify, in the direction you specify, with the stiffness you specify. For this reason it really only works well for coincident nodes. If you separate the nodes, the offset between them is not accounted for and you can see a moment imbalance between the applied loads and SPCs. A CBUSH helps to solve this problem because it does account for geometry (length). I think of a CBUSH as two rigid elements (like RBE2) joined at the middle by a CELAS with coincident nodes. I often model joints with the plates modeled at their mid-planes and the fasteners idealized as CBUSH elements with one stiffness defined idealizing the shear stiffness of the fastener. The moment due to the eccentricity gets transferred "rigidly", and doesn't disappear like it does with a CELAS element. There are some subtleties here, but especially for a joint with multiple fasteners, I don't think it should concern you too much that the plates are separated using CBUSH's. When in doubt, make a little test model to investigate the effects. That's what I always do!

 

[Naffoussi]

Hi,

Just a little thing about CBUSH elements. If the two nodes of the CBUSH are not coincident you will have a fatal error. So mke sure that your nodes are coincidents.

Regards.

 

[hezhj2000]

I don' think CBUSH requires the nodes to be coincident. I think CELAS requires coincident nodes.

 

[BlasMolero]

Hello!,
CBUSH is my preferred type of element, and not any problem exist for having non-coincident nodes. The CBUSH is a structural scalar element connecting two non-coincident grid points, or two coincident grid points, or one grid point with an associated PBUSH entry. As said above it has many advantages of using the CBUSH element over CELASi elements, for instance, if you use CELASi elements and the geometry isn’t aligned properly, internal constraints may be induced. The CBUSH element contains all the features of the CELASi elements plus it avoids the internal constraint problem.

In my opinion not any problem exist to have non-coincident grid points with CBUSH element. In any case I would suggest to make a pilot study and compare results, this is the way I run most of the times.
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/