Publication Date:
2004
Description:
Practical methods for the probability-based seismic assessment of structures
make use of estimates of demands produced by earthquakes of different intensities. The
uncertainties associated with these estimates are highly dependent on the variable
adopted as the intensity measure (IM, e.g., Peak ground acceleration, spectral
acceleration, etc.). This generates the need to compare the efficiency of an originally
adopted IMwith that of a new candidate. This implies comparing the dispersion of the
demand measure (DM, e.g., maximum interstorey drift ratio, ductility demand, etc.)
conditional to each of the two IMs. In order to obtain the demand estimates in a
conventional way, a full set of dynamic response analyses should be performed for each
IM under scrutiny, i.e., multiple records scaled at several fixed values of each IM. The
procedure developed here serves to accelerate this comparison avoiding the effort
required to evaluate the dynamic responses of the structure for all the ground motion
time histories considered every time that a new IM is adopted. For this purpose, use is
made of available results of analyses performed for a different (i.e., the original) IM.
Two methods are proposed: the direct method involves performing a regression of the
results obtained from the original analyses, taking the candidate IM as the independent
variable. The indirect method involves rebuilding the probability density function of
the DM given a defined value of the candidate IM by means of the total probability
theorem, using the results of the original analyses and certain data relating the two
IMs. The proposed methods have been tested by application to several SDOF systems with
different periods and different cyclic-response backbone curves. The conditions
affecting their approximation are explored, and some criteria to improve them are
identified. The procedure can also be used to determine the optimum value of a parameter
to be used in a parameter-based IM.
Keywords:
Earthquake engineering, engineering seismology
;
Non-linear effects
;
Dynamic
;
Intensity
;
performance-based
;
earthquake
;
engineering,
;
non-linear
;
dynamic
;
analysis
;
EESD
Permalink