Pressure Vessel Boundary Conditions

[DustinMechEng]

Hello Everyone-

I am using Cosmos/Works to analyze a sealed container that will be placed in a pressurized environment. I need to determine what the deflection of the container will be under a given pressure. My problem is that Cosmos requires me to fix some part of the assembly in order to analyze it, otherwise there is no boundary condition and the software cannot solve the analysis. However, fixing any part of the container keeps that part from deflecting and is not an accurate model of the situation. I tried the soft-spring setting, but the solver returned junk results. Any suggestions? Thanks for the help!

 

[Stringmaker]

Does your vessel have any symmetry planes? Also, models don't always have to be fixed to ground in order to prevent rigid body motion. If need be fix or constrain one single node in a particular direction to prevent rigid body motion. This can usually be avoided with some well thought out boundary conditions however.

-Brian

 

[jte]

Cross posted in thread1183-178236.

 

[corus]

Generally pressure vessels are axi-symmetric and you can use that to reduce your model down to 2D, otherwise use symmetry planes whereever possible as stringmaker suggests. Applying a restraint doesn't mean that you affect the deflection, at least not relative to cause stresses. If for instane you have a part heated up and restrain one end then the deflections will be relative to that restraint, but there won't be any stresses caused by the restraint, or at lest there shouldn't be if you've done it properly. You shouldn't need any soft srpings at all to restrain the part without affecting the deflection/stresses.

corus

 

[DustinMechEng]

Thanks for the suggestions. The problem is that the container is not symmetric about any of the center planes. In addition, essentially every node in the model will be displaced during loading as the container is sqeezed inward by the external pressure ... so contraining any part of the model results in an innacurate analysis. So far, I have contrained different parts of the model during several, separate analyses and "superimposed" the results. However, this isn't a perfect solution. Any other suggestions are greatly appreciated.

 

[johnhors]

You should be able to use the 3-2-1 method of supports as described in the recent thread:-

What is 'soft spring' thread727-176220

 

[corus]

In some instance a structure may not be truly symmetric but you can assume symmetry if parts are remote from others. An example may be nozzles from the vessel wall, which can be assessed as if part of a whole structure with symmetry at the cut planes. Some design guides on pressure vessels give the distances away from the nozzle to which you can impose symmetry. This may not apply if you're looking at buckling modes from an external pressure, as may be in your case. Otherwise use the method that john suggests.

corus

 

[DustinMechEng]

Thanks for the help everyone, I used the 3-2-1 method and it produced the results I was looking for!