I am experiencing difficulties understanding and passing the load cases around PSV's. First is I modeled everything properly with the guide and stopper location the hold-down with gaps etc. However, stress fails expansion case, in the reducer portion.

The way I understand it, due to the geometric figure of the reducer it is CAESAR II considers its thickness and dimater to the smaller portion that is why it fails, now as another method I tried modeling the reducer as pipe.


Diameter = Average of the two connecting pipes
Thickness = Average of the two connecting pipes

And now it passed the stresses. I am not sure if this is applicable and they also mention something like model it with a pipe but with an SIF, though this is the same thing with a reducer or is there any difference? Any comments will be appreciated. Thank you.

I don't know which piping code you're using, but in B31.3, reducers have a SIF = 1.

And without routing, there's not a whole lot of specific detail we can offer.

First line of thought:
Is the PSV to an open or closed system?
Which method did you use to calculate relief loads?
Are you over-constraining your pipe based on over-exaggerated relief loads, thus not providing flexibility for thermal?

Second line of thought:
Is the PSV small?
Some piping specifications only consider pressure when determining pipe thickness, which means small diameter piping become extra thin.
This results in you trying to balance loads on piping that's not much thicker than an aluminum can.

Third line of thought:
Are your displacements from your main line causing issues in your smaller branch line? Restrain the main line.

Fourth line of thought:
How stiff are your supports? CAESAR probably thinks it's 10e12 lb/in (or whatever it uses in SI notation). That's approximately the load to move one Pyramid of Giza 1/2". Perhaps you can estimate a more realistic stiffness.

To my knowledge, CII considers the section modulus of each end separately in calculating the stress. By modeling the reducer as one element with average thickness and diameter, you will miss calculating the stress at the weakest end of the reducer. A chain is only as strong as its weakest link, not as strong as the average strength of the links.

Faizal is correct, CAESAR II computes the stress at each end of the reducer using the properties associated with each end.