Buckling problem

[emeric33240]

Hello,

I working on a buckling case of head of  pressure vessel. I search the 1st mode of bucling.

I use different methods (shell, solid mesh, 1st or 2nd order, size of mesh), different meterial (Stainless steels of Invenor base or user base) and I find a value between 150 to 160 mbar. (-0.01 Mpa pressure, EIGV between 1.5 and 1.6)

My manufacturer say that this is generaly less than 50 mbar and this result is false.

I dont find the error or to justify the true in this case.

Can you help me ?

https://e1.pcloud.link/publink/show?code=VZK42QZTu1PjvaW1vQWUf9WKa4LIyVNYnG7

Regards

Emeric

[John_Holtz · Accepted solution]

Hi Emeric.

I had some problems with converting MPa to bar 😞. I have it worked out now and just want to clarify the situation.

• Nastran is calculating the vessel will buckle at a vacuum of 150 to 160 mbar.
• The manufacturer says they test the vessel and find that it buckles at 50 mbar.

If my understanding is correct, it seems reasonable. I am sure that this 1.5 m diameter vessel is not manufactured to the same precision as the CAD model. For example, the forming may not have a consistent thickness around the circumference, or the curvature has some variation, or the welding of the nozzle distorts the shell, and so on. It is quite common that the actual buckling load is lower than the theoretical buckling load due to such imperfections in the real part. A factor of safety of 3 seems reasonable to me.

Of course, you want to make the loads and constraints as accurate to the physical test as possible. For example, I know they do not put a pressure or vacuum on the vessel with that 300 mm diameter hole at the end of the nozzle. 😁If the physical test has a flange on the end of the nozzle, the pressure needs to be included. If the physical test somehow holds the nozzle end so that the load on the "end" of the nozzle is insignificant, then you need a constraint to mimic the arrangement.  The same is true for the constraint on the 1.5 m diameter bottom. Does the constraint applied in the analysis represent the physical test? (The constraint applied in the shell model is free to move radially and rotate about all axes.)

John

[emeric33240]

Hi John,

Thanks for our reply.

Yes effectively, I forgot to add the caps action, but without significants changes on results.

After compare with other sources, a safety factor of 2.4 to 3 is currently used. It was also confirmed by other source.

In conclusion, my model and results are ok, I just need apply a safety factor of 3 to be in accordance with real world (geometry disturbances).

Best regards

Emeric