Sol 105 and Pressure

[catiaUSR]

Hey!

Easy question.

Say I have a cylinder that is being crushed by external pressure. The pressure is a pressure field. So maybe on one side has 0.7 MPa and another 0.9 MPa and so on. And this is implemented by Pload2's on all the elements.

If I run a Sol 105 linear buckling analysis, will it increase *all* the pressures linearly? So if I get a load factor of 2, can I be sure the buckling load is 1.4 MPa on one side and 1.8 MPa on the other?

Sounds pretty straightforward, I just want to double-check it isn't increasing one load more than another etc.

Thanks!

 

[sdm919]

Simple answer: Yes, the buckling load is when all the applied loads are scaled by the eigenvalue.

Further comments:
1) Be sure to look at the mode shapes to see if they make sense. Sometimes local modeling issues can cause fictitious results.

2) There are times when you want part of the load to remain constant. Think of a pressurized cylinder such as a commercial aircraft fuselage. The internal pressure is constant as the external loads causing the fuselage to bend vary. SOL 105 can handle this with the STATSUB card. Use one STATSUB to define the constant part of the load "preload" and one STATSUB to define the variable part "buckling". In this case, the eigenvalue scale factor only applies to the variable part. Read Nastran documentation for details.

3) Remember that the SOL 105 solution is a linear eigenvalue solution. So for metals, it does not account for plasticity, so check that the buckling stresses do not exceed the yield stress anywhere. If they do, the eigenvalue solution may be over-estimating the buckling.

4) SOL 105 also does not account for initial imperfections in the shape of the structure, which can be important in thin shells. If a cylinder with external crushing pressure is very thin and not exactly circular (even though the FE model is), the real structure may collapse sooner than the eigenvalue solution would predict. That is why thin shells often use "knockdown factors" that are applied to theoretical solutions to better agree with experiments. One source of these factors is NASA-SP-8007 Buckling of Thin Walled Circular Cylinders, which is available on the NASA web site.

 

[catiaUSR]

Thanks sdm

Those are all good points. Yes, I am familiar with STATSUB and very familiar with SP-8007.

I'm totally on board for statsub with skins that take fuselage and pressure loads, and pressurized tanks that take beam loads but what about fairings.. If the only load (outside of handling/ground loads) on a part is an aeroload, and buckling sizes the part, then I feel like this is a genuine case of the pressure needing to be stepped up.