Need a book? Engineering books recommendations...

Return to index: [Subject] [Thread] [Date] [Author]

RE: Floor response spectrum calculation

[Subject Prev][Subject Next][Thread Prev][Thread Next]
You are jumping right into the middle of it. 
The big issue is if you choose to ignore vertical ground motions.  Essentially the current ASCE 7 drives you to a default that is 20% of the horizontal for the vertical with no mention of time domain.  If you ignore vertical, you can use more methods like the spectra to spectra method.  But it is conservative. 
If you need to consider vertical (I would always consider vertical), you need to do a time-history analysis.  That would drive you to use a tool like SAP.  You input the time-histories and place a reasonable oscillator at discrete points on the structure and capture enveloped response spectra. 
What you are ultimately driving toward is what is required for an in-structure response that is used for equipment seismic qualification.  ASCE 4 has a lot of good information on this.  Once the in-structure response spectra are developed, they can be compared to a fragility curve for equipment or you can use that data to design component supports. 
ASCE 4 has some information as does DOE EH 0545 and the UFC 3-310-04.  You might also want to evaluate kinematic soil structure interaction (KSSI) if you are going down this road. 

Regards, Harold Sprague


From: bauskas(--nospam--at)
Date: Tue, 17 May 2011 10:32:26 +0300
Subject: Floor response spectrum calculation
To: seaint(--nospam--at)

Now we have the two universally accepted methods to do dynamic analysis of a structure that both rely on modal decomposition: spectral method and direct dynamic method (or whatever you call it, time-history or transient).

If there is a task to calculate the response spectrum at a given node of the structure (over frequency domain) from a given ground motion record (accelerogram), do I get it right that the only way to do this is to perform a full-blown time-history analysis?
The logic behind this is, to calculate the spectrum you have to have the actual response record at the desired node, and the spectral approach won't produce this kind of information. Is this right?

Any reference to a textbook would be a very nice thing to have. Many thanks in advance and sorry that I have flooded the list with my threads.

P.s. If you ever need to convert an acceleration record from one text format to another, drop me a note. I have programmed a _vbscript_ to do this for our local format and hopefully will be able to tailor it to another format.
I even have a code to produce a response spectrum from any amount of time-history earthquake data (still testing this though).

Alexander Bausk
Civil/Structural design & inspection engineer, CAD professional
MSc Structural engineering, Ph.C. Engineering
Dnipropetrovsk, Ukraine
Tel. +38 068 4079692
Fax. +38 0562 470263