Dear All,

Today i was analyzing a system in which 2 of the nozzles are connected to a vessel. The vessel nozzles are of Elbow type. I modeled this nozzle as a elbow nozzle in the Caesar with the vessel material & ran the analysis. As a result my nozzle loads were very low. Later i changed the elbow nozzle to rigid component and then ran the analysis & i get all nozzles exceeding.

Why is this happening. Advise me which method of analysis is correct. I am attaching the image of the system for your reference.

"Later i changed the elbow nozzle to rigid component and then ran the analysis & i get all nozzles exceeding."

By modelling the component as a rigid you are increasing the stiffness by a factor of 10. Flexibility at the vessel wall is useful as this reduces the loads on the nozzle. Why not model it as pipe with the correct thickness?

Perseus,

I am not able to get " Why not model it as pipe with the correct thickness?",
Does it mean that i shall model the vessel also???

I was implying that you can model the nozzle wall thickness (with its reinforcement). It looks to me like you are modelling the vessel already.
However, if using rigids to model the vessel is confusing the matter then replace the vessel elements with a displacement for vessel expansion. Compare the results and evaluate whether you were doing this correctly to begin with.

If I understand the original question correctly the main issue is that you are replacing a flexible component (elbow) with a rigid. Thus whatever happens at the 'from node' more or less happens at the 'to node' in terms of displacement.

Can I suggest that the reducers might be eccentric btw.

Perseus,

The elbow is actually a nozzle & the wall thickness is as per vessel calculation.

My intention of modeling the vessel is for thermal expansion. Moreover every nozzle has some flexibility, since in my case its elbow nozzle i just modeled the same instead of a rigid component(i.e. my consideration is nozzle being flexible & vessel being rigid)

Yes, you are right the reducer has to be of eccentric type with flat bottom.

Couple of comments:

Using rigids will make the system conservative from a piping and nozzle standpoint, and if you're checking the reaction forces on the nozzles, they will be conservatively high.

Typically, I model my vessels, and they are anchored down.

If using ASME B31.3 or B31.8, they do not address eccentric reducers. You can, however, model it in CAESAR II if you give it the appropriate vertical displacement. For stress analysis purposes, I typically double the SIF, as that follows the practices of ASME VIII with respect to eccentric conic frustums.

An elbow is still an elbow whether it is part of the piping or supplied with the equipment. The "equipment" elbow should be modeled; use the OD and T from the equipment dwgs/data.

The CII model is a mathematical representation of a physical piping system. Represent it. Of course a rigid is "typically" more conservative for EXP but not necessarily for SUS/OCC and the overall load distribution may shift in the model.