RSA and Time-History Analysis 2

[humanengr]

I'm using STAAD to perfrom RSA analysis of multi-level steel structure. Structure is industrial and mostly supports equipment and vessels as well as some working platforms.
I'm trying to determine if Time-History Analysis is necessary or if RSA is sufficient.
If anyone has some feedback on that, it would be appreciated.
Also if anyone has any thoughts on whether it is necessary to model the foundation with the steel structure, that would be helpful. Structure has five levels, some horizontal irregularities. Foundation will be mat foundation on driven piles. Located in "high" seismic region.

 

[TERIO]

In my understanding, a Response Spectrum Analysis will typically give a more conservative result than a Time-History Analysis (assuming that the response spectrum of the time-history matches the design response spectrum reasonably). The trend I see is to move from equivalent static to response spectrum to time history, in that order, when trying to reduce conservatism in the analysis.

 

[jdgengineer]

RSA analysis and Time History Analysis are both considered Linear Dynamic Procedures and as far as I know ether is allowed if linear dynamic procedures are allowed.

Whether linear dynamic analysis is allowed will be dependent on the code you are designing to. Per Table 12.6-1 of ASCE 7-05 RSA analysis is allowed for all structures. However, existing structures analyzed to ASCE 41 have different requirements and there are limitations to the usage of linear analysis techniques (based on irregularities and DCR) that require the usage of nonlinear techniques.

RSA is using statistical methods (SRSS or CQC) to determine the peak response of the building. Time history however measures the peak response directly. I seem to remember from my graduate school days that for a SDOF structure the results should be the same. The approximations of the RSA method come from the modal combinations.

I believe if you are very near to the fault that time history analysis can better capture some of the near source effects. However, if that is the case nonlinear time history analysis may be mot appropriate.