Advances in Seismic Design Methodologies - Earthquake Engineering for Structural Design - page 435

Advances in Seismic Design Methodologies 435

structures, but the currently available computer hardware and design software
effectively limit both size and complexity of structures, which may be analyzed using
this procedure. At present, there is no general purpose non-linear analysis software
which permits practical evaluation of large structures (Lee, 2008). There are three
barriers preventing the engineer from the application of non-linear dynamic earthquake
analysis:
- It is too complex to be solved by means of personal computers, currently used in
structural practice.
- The results are too sensitive to the properties of structural material non-linearities.
- The analysis results significantly vary in function of the earthquake records used.
Therefore, until these barriers will be eliminated, this procedure remains a very
useful tool for research works only. The main problems of using this methodology are
related to times history representation of ground motion modeling, hysteretic models
for structural material and element behavior, time-history analysis and incremental
dynamic analysis.
Time-history representation of ground motions. The ground motions are very
difficult to predict in terms of time of occurrence, intensity and duration. The definition
of spatial characteristics, like acceleration velocity and displacement in three
directions, together with the temporal ones, like period and duration, is also
considerably difficult. With respect to Earthquake Engineering, there are two ways in
which ground motions can be represented (Gioncu and Mazzolani, 2002):
(i) Using recorded digitized time-history representation from a data bank,

classified in function of source type, distance of the source-recording station
and station soil characteristics. So, the designer can select the most
representative time-history records, corresponding to the involved site. This
way presents many difficulties, due to the fact that the codes ask for the use
of a minimum of 5 records for the structure analysis. In spite of the large
number of available records, it is very difficult to have a sufficient number of
records with characteristics similar to those expected at the site of the
structure. On the other hand, the large variability in the characteristics of
ground motions, like peaks of acceleration or velocity, period, duration, etc.,
which are able to influence in very large measure the structure response,
produces an important scattering, without the certainty that the structural
response to selected ground motions represents the actual one.

(ii) Using artificially generated time-history representations is a very convenient

way to solve the above-mentioned problems. These ones can be created by
the combination of actual recording portions. The artificially generated
representations are not necessarily the representation of a physically realistic
earthquake; they should be considered as encompassing all possible seismic
loadings, corresponding to the source type, dominant period, soil conditions