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

RE: UBC 1630, Ev Question

[Subject Prev][Subject Next][Thread Prev][Thread Next]

There is an issue that I looked into when I was working on a case study of a
building that was damaged in the Northridge Earthquake.  Just after the
earthquake, a number of people were taking about the possiblity of vertical
accelerations having contributed to damage.  When I looked at the issue, it
seemed pretty straightforeward to break down the issue, as follows:

First, geotechnical engineers and engineering geologists say that the vertical
response spectrum in the immediate near field of the event tends to be about
equal to the horizontal spectrum.  Further away, it is commenly taken as 2/3 of
the horizontal spectrum..  Taking a typical zone 4 spectrum, the peak ground
acceleration is 0.4 g.  The peak amplified response is 2.5 x 0.4 = 1.0 g.  If I
take the worst case of response for any structure (i.e., the peak of the
spectrum), then, the vertical response in the near field might be as much as 1.0
g, and further away, 0.67 g.

Second, the peak response from vertical motions is very unlikely to occur in the
same instant of time as the peak horizontal force.  So peak vertical forces
should not be directly added to peak horizontal forces. Instead, a statistical
combination is needed.  Using the UBC as a basis, the likely maximum of vertical
acceration that might occur in the same instant as the peak horizontal
acceration is 30 percent.  Thus, in the near field, as much as 0.3 g might be
considered in conjunction with design horizontal force, while further away, 0.3
x 0.67 = 0.2 g might be used.

Now, in term of my personal opinions, the above analysis is pretty simplified
and conservative, since it assumes that the vertical response occurs in a single
mode, which has a period that is in the peak range of the spectrum (vertical
mode shapes tend to consist of a large number of modes of small areas of the
structure and have very short periods, closer to the zero-period acceleration).
So I think that design provisions could be a bit more permissive.  So in my
personal opinion, I think that the current use of 0.9D (or 0.85D in the ASD
combinations) is adequate for areas not in the near-field region (say, when the
near-source factor Na is 1.0), but that there might be some justification for
using a lower coefficient such as 0.8D or 0.7D when located in a near-source