Dear All,

I’m performing a dynamic earthquake analysis considering different response spectrums for each building´s level. Pipe stresses exceed admissible values by far in this case, but if I apply the most severe spectrum (buildings highest) uniformly to all piping system stresses result admissible. It seems something is going on in the levels transitions, is there any special care to take in a multi-level spectrum analysis?

Thanks

When more than one spectrum is applied, there will most-likely be a relative displacement component to the load. (One structure or floor may be wagging left while the other is wagging right.)
This displacement set may be the cause of your high stress. If it is the largest of all stress components, it will be indicated by "P" (for pseudostatic) in the stress report.
You specify this relative displacement in the last column of the spectrum definition. If you do not specify a value, CAESAR II will calculate the value from the response spectrum. In many cases the values extracted from the response spectrum are excessive.
(Check your seismic displacements.)

Hi Dave, thanks for the prompt answer.

I’m wondering now hoy to determine appropriate displacements for each node… do you have any method?

Would I be in the safe side if I apply the most severe spectrum to the whole piping system?

Thanks!
Martin

I would think the same method used to collect each floor's response spectrum could also be employed to develop the floor's "zero frequency" deflection. (CAESAR II uses the acceleration of the lowest spectrum frequency calculate this displacement - displacement=acceleration*omega squared, where omega is that lowest frequency in radians.)
Enveloping all the floors into a single response spectrum will eliminate these pseudo-static floor displacements but I cannot say that this is conservative or appropriate.
Also, this displacement is not applied to each node. It is applied to all boundary conditions (e.g., anchors, restraints, displacement sets, hangers, etc.) on that floor.

Ok, thanks Dave!
Best Regards
Martin

hi,
i would like to highlight the following points in seismic analysis based on my experience.

1)Generally multi point response spectrum should yield low stress results,compared to envelope spectrum.)

2) IF you are going to consider seismic anchor movement(SAM),it induces secondary stress so allowable limit is very high compared to inertia loads.(check the code allowable limit for SAM)

3) Do multi point response spectrum analysis by setting SAM as Zero and check the results,if still exceeds the limit compared to enveloped spectrum kindly check the grouping of spectrum .Try a piping benchmark problem in Nureg report (where they have compared results for envelope and multi point response spectrum method for identical model)

regards
Udayakumar

Thanks Udayakumar! I will look for the Nureg report, remember the name?
Regards
Martin

You can see this in the Example job NUREG9, located in \Examples.