Nastran output on Shell and theory

[sushi75]

Hi everybody,

Just a quick question about stresses on shell elements.

I've a got a FE model of a tank, sujected to internal pressure, and I looked at one elements output results.

I wanted to check that given the force and bending moment, I should by hand calc have the same value of the stress from Nastran output, according to sigma=F/S+-My/I

Im aware that membrane forces and bending moments given in Nastran are per unit length, but when I do the calc, the value is not exactly the stresses in the output.

Does this discrepancy comes from the shear force that is not taken into account in the formula? If so, how can I include it?

Hope you can bring some some guidance on that!

Thanks a lot,

Cheers

 

[rb1957]

how does shear stress contribute to normal stress ?

you say the hand calc does not exactly agree with the FEA. how much "not exactly" ? 99%, 95%, 90% ... ?

are you looking at a flat plate under pressure ? then FEA is probably having issues with large displacements.

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

 

[sushi75]

Hi,

Thanks for your quick reply. Actually, there is a good 10-15% diiference between calculated stress from force and moments and the direct stress output.

The pressure load case is high thus stress values are also high. But still does that explain such a difference?

Unless there is a more accurate formula for plate which I'm not aware of??

Cheers,

 

[sushi75]

It works now! I have checked the element thickness and it's smaller than the one specified in the element property card.

So with the actual value, stresses are ok!

 

[rb1957]

? do you mean you've checked the model against the design, your calc using design data vs FEM calcs, and found that the plate thickness by design (as used in your calcs) was different to the FEM element thickness ? ... which would make perfect sense !?

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