Civil Engineering Reference
In-Depth Information
appropriate selection parameters depend on which characteristics of the selected motion are considered
most important from a structural response viewpoint.
The selection process is also a function of the objective of assembling a strong- motion record suite.
For example, if the records are required for a specifi c site subjected to a well- defi ned hazard, normally
characterized by magnitude, distance and site condition, the selection process would be distinct from
the case where an engineering fi rm requires a number of records to be used in routine analysis of a
wide variety of structures of yet- to - be - defi ned characteristics. Assuming that the two examples repre-
sented above cover many application examples for structural earthquake engineers, simple procedures
can be defi ned for each as discussed hereafter.
(i) Matching a Design Scenario
Matching a design scenario often corresponds to reconciling magnitude- distance - soil condition triads.
There are, however, uncertainties in magnitude and distance calculations. Therefore, it is reasonable to
select records within ranges from the design event to increase the possibility of fi nding a viable record
suite. Guidelines are lacking but searching within a range of ±0.3 magnitude unit and ± 20 - 40 km dis-
tance is reasonable. Databanks of well-constrained records are examined, e.g. those presented in Section
3.7, provided the origin of the magnitude and distance calculation, as well as the site condition catego-
rization, are known and accepted and a number of records within this range are selected. Matching the
magnitude is more important than matching the distance. This is supported by the discussion on elastic
spectra of Section 3.4.2. For inelastic cases, the increase in distance would lead to an increase in dura-
tion, which will in turn affect the response. Thus, the records obtained are unlikely to have the design
PGA. Scaling, using a recommended procedure as illustrated in Section 3.5.4, should then be applied
to arrive at a set of records that will consistently test the structures intended for the site. This is one of
the simplest procedures for selecting records to fulfi l a predetermined design scenario.
(ii) A Suite of Records for Design Offi ce Applications
Whereas this is not a technically robust requirement, it is often requested by the non- specialist practice.
For example, it is suitable for design offi ce use, where limited knowledge of earthquake engineering
and engineering seismology exists. Also, it is a useful approach for investigating seismic response in
non - specifi c region applications (e.g. Broderick and Elnashai, 1996, among others).
Records are selected on the basis of the peak ground acceleration (PGA) to the peak ground velocity
(PGV), i.e. the ratio PGA/PGV (Zhu
et al.
, 1988). The rationale behind this is that near-source shallow
earthquakes or records measured on rock, will exhibit high acceleration peaks of short duration, leading
to low-velocity cycles. These records will give high values of PGA/PGV. Deep or distant earthquakes
or records measured on soft ground will have lower acceleration values, but individual cycles are of
longer duration, leading to high-velocity waves. These will yield low PGA/PGV ratios. Intermediate
scenarios in both senses will yield intermediate values of PGA/PGV.
With regard to structural response, high values of PGA/PGV records will be more critical for stiffer
structures, while more fl exible structures will be strongly shaken by low PGA/PGV records. Therefore,
selection of records based on PGA/PGV, with a reasonable number in each of the regions low, medium
and high, will ensure that the ensemble is capable of imposing high demands on structures in a wide
range of periods and will implicitly include many of the engineering seismology features related to
source characteristics, travel path and site conditions, as mentioned above. The approximate ranges of
PGA/PGV ratios determining the low, medium and high ranges are as follows:
low
PGA PGV
<
0.8
(3.29.1)
medium
0.8
≤
PGA PGV
≤
1.2
(3.29.2)
high
(3.29.3)
1.2
<
PGA PGV
