Civil Engineering Reference
In-Depth Information
3.5.4.1 Scaling Based on Peak Ground Parameters
The loads acting on a structure during an earthquake are proportional to its instantaneous acceleration
due to the imposed base motions. On account of this, recorded ground motions are conventionally
scaled to a PGA value prior to their application in response history analysis. This method of scaling
has advantages in that it is simple to apply and agrees with the methods through which design codes
normally defi ne seismic loads. However, PGV and PGD values also play a signifi cant role in determin-
ing the severity of seismic response.
In general, it is possible to identify three ranges of structural periods within which the response is
dependent on the values of ground-motion acceleration, velocity or displacement, as also illustrated in
Section 3.4.4. Short-period structures (typically less than 0.5 second) are sensitive to peak ground
acceleration, while structures of moderately long period (i.e. 0.5 to ∼2.0 seconds) are sensitive to peak
ground velocity. The response of structures of exceptionally long period (i.e. longer than 2- 3 seconds)
is likely to be more dependent on displacement. The dependence of intermediate period structures on
velocity is directly recognized by some codes of practice, such as in the Japanese code (BSL, 2004 ),
where peak ground velocity is used for the design of tall buildings. Therefore, it is appropriate to scale
earthquake records in a manner that refl ects the response periods of the structure under consideration.
This has been confi rmed by a number of studies on large suites of earthquake records (Chandler, 1991 ;
Tso
et al.
, 1992) in which the degree of spectral dispersion is reduced in the low- period range when
acceleration scaling is applied and a similar reduction is observed at longer periods with velocity
scaling. This observation applies equally to the important parameter of the ductility demand imposed
on structural systems when the additional effects of varying yield strength must be considered.
While replacing PGA with PGV for intermediate period structures offers an improvement over using
acceleration for the full range of period, it is still not suffi ciently accurate. It was shown by several
researchers (e.g. Nau and Hall, 1984; Matsumura, 1992) that PGA and PGV are not always adequate
measures over the wide range of frequencies, since they are based on a single point of the response
spectrum. Response velocity spectra of various ground motions normalized by PGV may be remarkably
different. Furthermore, existing codes of practice require seismic loading to be uniquely defi ned in
terms of PGA, with an associated acceleration response spectrum. This implies that scaling ground
motion to a common peak velocity would disturb the equivalence between the time- history record and
the design spectrum.
3.5.4.2 Scaling Based on Spectrum Intensity
In view of the above, spectrum intensity, rather than a peak ground parameter, may be utilized as a
reference for scaling earthquake records. This scaling procedure assumes that the seismic energy
imparted by the scaled earthquake record is equal to that implied in the design spectrum of the adopted
seismic code. Several spectrum intensity scales have been suggested in the literature, the most pertinent
of which are discussed hereafter.
(i) Housner Spectrum Intensity
The spectrum intensity proposed by Housner (1952) has received considerable attention in conjunction
with scaling procedures for earthquake records. Using the assumption that elastic response spectra may
be used to estimate the energy available to cause damage, it was suggested that the velocity spectrum
could provide a measure of the severity of structural response. The intensity of shaking of an earthquake
at a given site was represented by the spectrum intensity
SI
H
, defi ned as the area under the elastic
velocity spectrum, between the periods 0.1 and 2.5 seconds:
2.5
∫
SI
=
S
(,)d
T
ξ
T
(3.30)
H
v
0.1
