Need a book? Engineering books recommendations...

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

Fw: Linear Time history Analysis against Response spectrum Aalysis

[Subject Prev][Subject Next][Thread Prev][Thread Next]
I would have to agree with the two points listed below.  Both an elastic TH
and an elastic RSA are...ELASTIC.  Unless one is designing a nuclear power
plant, the seismic design will usually not be an elastic one.  Therefore,
an ELASTIC analysis is of limited value, and more emphasis needs to be
placed on designing the structure for INELASTIC action.  Elastic time
history analyses have always seemed a strange animal to me, as (most)
structures are not expected to endure an entire design level earthquake
without yielding, and the first instance of significant yielding throws the
TH analysis out the window, so why do it?  The same could be said, to a
certain degree, for elastic RSAs.

Designing a transfer structure (which our office has done for a 13 storey
building) SHOULD be done in such a manner that the transfer element can
sustain the maximum loads that the structure above is capable of putting on
it, including the effects of material overstrength and strain hardening. 
This is really the only sound method for "assuring" that the elements which
are intended to dissipate energy (via plastic deformation) will actually
get to do their job without some other element (most likely the transfer
element) failing first.

The best approach, IMO, would be to start with a basic equivalent static
force (first period) analysis.  Using a static force approach you will
still have equilibrium (and signs), unlike an RSA.  Do a basic design of
the upper portion of the structure for the reduced design level forces. 
Then, calculate the likely capacities of the upper level structure.  This
will give you likely maximums (based on member capacities, NOT some
fabricated ground motion or spectra) which you can use to design the
transfer element.

IMO, this approach makes the most sense because it places much less
emphasis on the ground motion modeling, which is basically a guess in most
areas that don't have the benefit of site specific seismological data. 

After doing a capacity based design, you can run all the time histories OR
RSA's that you feel are necessary to validate the results (and satisfy code

My .02 EGP

T. Eric Gillham PE

> To recap my points:
> 1) You have to move beyond an elastic analysis when dealing with a
> structure 
> subject to seismic loading.
> 2) An elastic time history analysis is not necessarily any better than a
> response spectrum
> analysis.
> Mark Gilligan
> Markkgilligan(--nospam--at)