On seismic behavior and vulnerability of building structures: Dynamic identification, condition assessment, and fragility modeling

Abstract

Structures undergo various changes during their service life due to deterioration or external loading. These changes are the keys to design, assess, and predict structural behavior. The three pillars of structural engineering are the most governing factors to justify safety, serviceability, functionality, and operability of structures, considering forms and functions. Although independently extensively studied, coordinated efforts have not been widely practiced in structural engineering, starting from characterization to condition assessment and finally probabilistic response depiction. Aiming at coordinated dissemination of the three pillars, the present study performs system identification based modal characterization to develop period-height relationships and condition assessment under no external loading, structure undergoing deterioration, changed state of structure due to extreme loading, changed state of structure due to construction progression, and strengthened structures following various states of damage. Furthermore, many sets of seismic fragility functions have been developed for low-rise RC buildings, considering global and component level performance using empirical data. This study covers RC and stone and brick masonry buildings for dynamic identification, condition assessment, and seismic vulnerability modeling. Experimental and empirical approaches have been deployed to estimate dynamic characteristics as structural performance signatures, evolution of dynamic characteristics as signatures of changed state of the structures, and probabilistic models as signatures to quantify the exceedance probability of particular damage level in buildings and components under seismic excitation. Bayesian modeling based formulation is adopted through MCMC using Metropolis-Hastings sampling to derive period-height relationships and fragility functions. Subsequently, posterior parameter uncertainties have also been characterized.