| dc.description.abstract | This dissertation presents an original investigation of several topics related to seismic
analysis and currently-implemented code-design procedures for seismically isolated
structures, focusing primarily on high-damping rubber-based isolators.
A considerable effort to develop accurate mathematical models to represent rubber-based
isolators' force-displacement behavior has been made in the last decades. These researchoriented
models can represent complex phenomena as shear-strain hardening, scragging, and
strain-rate dependence. However, engineering design procedures have not embraced these
advanced modeling techniques, and the implementation of equivalent linear or bilinear
models using deterministic parameters is still recommended for seismic response
assessment. This approach neglects the model parameters' inherent uncertainty and ignores
several characteristics of the isolators' force-displacement relationship, whose relevance
should be elucidated.
As the number of isolated structures increases steadily in Chile, other Latin American
countries, and most of the world's seismically active regions, this research aims to close
some aspects of the gap between the research-oriented modeling techniques and the
simplified engineering design procedures. Particular emphasis is placed on the uncertainty quantification of the isolators' effective properties currently used in engineering design
procedures. To reach this goal, this thesis is divided into three stages: (i) the development
of a simplified and versatile element model for seismic isolators' response history analysis,
to be implemented in engineering design practices but able to capture accurately relevant
features of isolators' behavior; (ii) an uncertainty analysis of the properties used in equivalent
lateral force and response spectrum procedures, quantifying the variability of the measuredby-
test effective stiffness and effective damping of a vast isolator dataset; and (iii) a
statistical analysis of damping modification factors used to correct the seismic demand in
equivalent lateral force and response spectrum procedures, aiming to find better predictors
for these damping factors based on spectral shape metrics.
It is highly expected that some findings of this research can improve current design
methodologies, allowing for a better estimation of interstory drifts, inertial forces, and floor
accelerations on protected structures at a reasonable additional effort. The implementation
of the element model presented in this thesis in general-purpose software-packages for
seismic analysis is encouraged. Moreover, some findings related to effective properties
uncertainties and damping modification factors could enrich current provisions in the
Chilean design code for seismically isolated structures. | |
