Bolt connection 2

[Venkat133]

Hi sir
I have a small doubt in nx nastran. i.e when I connect two components with bolt connection with spider elements iam getting huge stress on spider connect nodes because of shear load. For same model when I take glue connection stress are very less(because two components are connected with more number of bolts). Even I take avg stress one element away from red zone ( st. Venants principle) which stress I should consider

 

[DraftingMan]

Venkat133,

You will get a better answer in the nastran forum.

John H. Dunten, CD
Certified Drafter

 

[EnginerdNate]

John is right, this would be better answered in the nastran forum.

That said, stresses at the connections to a spider element are never accurate. Are you including any compliance in the form of a spring element between your spider elements or are you simply tying your two bolted layers together hard with an RBE? If you're doing the latter you'll always get huge stress concentrations (at least with RBE2s) because you're adding an enormous amount of stiffness (infinite) between those two parts of the joint.

Typically if you need to model a bolt as an RBE spider + a spring (CBUSH) connection, you'd ignore stresses in the elements directly connected to the CBUSH (at least for gross acreage stress/strain checks) and calculate your margins at each bolt location by hand based on the forces being transmitted through the CBUSH, checking against fastener strengths (tensile/shear) and parent material/joint allowables (bearing, fastener head pull through, etc).

When you setup your glued connection, what elements are you glueing?

Cheers,
Nathan

 

[rb1957]

we seem to have forgotten that every stress/number from a linear FEA is not some drop of truth from the god(s).

Any stress above yield in a linear FEA is "fiction".

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

 

[SWComposites]

And people STILL don’t understand that virtually every stress in an element connected to a rigid element is rubbish.

 

[Mohanlal0488]

Just my 2cents.

When dealing with with FEA's for connections I still use structural codes for checks around the bolts and areas where the results may not be accurate.

 

[EnginerdNate]

I should have caveated my answer with the fact that I generally deal with materials where if anything is truly pushed past the elastic regime the goose is already cooked.