Extracting bolt loads from cbush elements

[Loc13]

Hi everyone,

I'm trying to model a plate with two c-channels bolted underneath it in Patran/Nastran. The plates and c-channels are modeled with shell elements, and the bolts are modeled using rbe2 with bushing properties in order to extract the bushing forces for bolt analysis. Based on my hand calculation for the shearing force on the bolt, the bushing forces (shearing component) I extract from my model are significantly higher than what my hand calculation suggests. I think this is most likely due to the stiffness values I assign to the bushing property, calculated using K=AE/L.

Is K=AE/L the correct way to determine the fastener stiffness value for bushing elements? Does this K value also apply to rotational stiffness? In general is this a good way to extract bolt loads?

Thank you very much in advance!

regards,

 

[40818]

you should determine the fastener spring stiffnesses of the bolts. The "Huth" method is probably the most accurate/preffered option, but there are other methods available, (Tate & Rosenfeld, Boeing, Schwarmann, Swift etc). The flexibility of a fastener is generally defined as the deflection of the joint around the fastener, no tincluding th enormal extension of the material. There are so many variables which affect the joint stiffness and even correlating back to test data is sometimes tenious (so you could use Huth's method for example and apply it to very thin sheets and get crap results as his work was based on tests of thick sheets. The only way you can get accurate stiffnesses is to test, but that isn't realistically practical.
My preferred method is to use 3 celas elements for each bolt (2 shears and a tension), and use Huth to obtain the stiffnesses for the FE.
The saving grace is that even when you use one of the methods above, in the great scheme of things it makes little difference.

 

[Loc13]

Thanks for the reply 40818, I will look into the methods you suggested!

 

[jetmaker]

Trying to find information on the Huth method... can someone provide some links to papers, etc...

jetmaker

 

[SWComposites]

Search the Aircraft_Eng and Aero_Eng_Other_Topics fora for several discussions of fastener flexibility methods, including Huth.

 

[darkwing888]

Instead of a RBE2 element, maybe a bar element would give you reasonable results.

"Clever of me to use my spine to break my fall like that."

 

[Loc13]

I've been looking into modeling fasteners with CFAST or CWELD and trying to do some case studies with these two types of connectors. Can anyone comment on their experience with these two particular connectors/elements?

 

[40818]

Darkwing;

"Instead of a RBE2 element, maybe a bar element would give you reasonable results"

I dont think so, with a bar element you can enter values of shear stifness for the elemental Y and Z axis, but i dont know how you could enter a value for the X axis.

 

[darkwing888]

40818 - axial stiffness = AE, bending stiffness = EI, torsional stiffness = GJ, etc. With NASTRAN you can define a CBAR element with all these stiffness values as well as offsets and end releases.

"Sheesh. Some days it just doesn't pay to get out of bed."

 

[40818]

darkwing888;

The standard stiffness calcs you mentioned are usual for comparing beams etc, but bolt joint stifnesses are far more complicated. Compare specialised bolt stiffness theories and test results, against the standard stiffnesses and see how they compare.
I'm not saying using a bar element is bad, just that using celas's is better. Though the main thing is to get decent elements to represent the actual bolts and joints, anywhere within 10-15% and its going well!!