Challenges for the Next Generation of Seismic Codes - Earthquake Engineering for Structural Design - page 469

Kocaeli (Turkey), yielded more near-source records. But, unfortunately, there
is still a large number of seismic zones without adequate information.

- Due to the unknown fault network (e.g. Kobe earthquake, produced along an

unknown fault) or some active faults different to the known ones (see
Northridge earthquake produced by Elysian Park fault far from the very well-
known San Andreas fault), it is practically impossible to design recording
network stations in a such a way as to have records on the sites where the
phenomenon of polarization shows the way to the maximum ground motion.
So, the recorded values have a great incertitude and, only by chance, some of
them are situated near the fault (e.g. Northridge, where, in spite of a large
amount of recorded data, a very reduced number of records is available for
near-source ground motions). In addition, these records can be influenced by
some local site stratifications, so that they are available only for the recording
station (see the 1977 Vrancea ground motions, recorded in Bucharest on a very
poor soil condition, which led to the wrong conclusion that this source is
characterized by long natural periods). Therefore, the design spectra developed
based on these recorded motions are not, generally, available for other seismic
sites.

- During recent decades, one can see a tremendous growth of urban areas in

seismically active regions, which increases the risk that an earthquake occurs
where there is a large concentration of population. The damage produced in
these cases is well illustrated by the above-mentioned earthquakes. So, the so-
called near-source earthquake must be taken into account in structural design,
but these aspects are not considered by the response spectra used in practice.
Only USA-UBC 97 and Chinese GBJ11-2001consider some aspects of the
near-source earthquakes. The UBC proposes an increasing of acceleration
values in Californian near-source zones, in function of the distance from the
fault (but who knows exactly this distance?). The GBJ11-2001 proposes
different corner periods for near- and far-source earthquakes. However, the
increasing of acceleration or reduced corner periods are not sufficient
measures to consider all the very damaging effects of near-source earthquakes.

- Peak ground acceleration, in the frame of the used design spectra, is the

parameter associated with the severity of an earthquake. However, it has been
generally recognized that it is a poor parameter for evaluating the earthquake
damaging potential. For instance, one of the more significant shortcomings of
the current design spectra is the fact they do not account for the duration of the
input ground motions. The dissipation of seismic energy, as it is considered in
the design philosophy today, is strongly affected by this duration. The near-
source impulse type of ground motions (in case of crustal earthquakes) results
in a sudden insertion of energy into the structure, which must be dissipated
immediately by very few large yield excursions. Contrary, for long duration
earthquakes (in case of far-source or deep earthquakes), the cyclical type
ground motions with numerous yield reversals require a more steady
dissipation of energy over a long period of time. The actual design spectra do
not take into account this difference. So, the question is whether the structures
are able to face these different situations.