Midway evaluation in Civil Engineering - Dipendra Gautam

Title of thesis: On seismic behavior and vulnerability of building structures: Dynamic characterization, condition assessment, and fragility modeling

Student: Dipendra Gautam

Doctoral committee:
Dr. Rajesh Rupakhety, Professor, Civil and Environmental Engineering, University of Icealnd
Dr. Bjarni Bessason, Professor, Civil and Environmental Engineering, University of Icealnd
Dr. Hugo Rodrigues, Associate Professor, University of Aveiro

Abstract

Seismic performance of a building is dominantly characterized by its stiffness, which also governs both design and assessment at any stage. Stiffness signatures are primarily expressed in terms of vibration frequencies and damping ratios. While dynamic characterization is extensively reported in the literature, large scale experimental campaigns and pragmatic formulations such as period-height relationships are rare. Furthermore, evolution of dynamic characteristics under changed states, damaged to retrofitted, are rare too. Both aspects deserve further attention to better quantify seismic behavior of existing buildings. Predictive models that outline the probabilistic occurrence or exceedance of particular damage level under the given dynamic loading are backbone of performance-based design and assessment. For this, seismic fragility models are required. The vast majority of the existing fragility models do not account for uncertainty quantification although uncertainties are inevitable due to widespread variability in material, geometry, construction practices, among others. To address these challenges pertaining to seismic behavior and vulnerability of building structures, this study deploys experimental and empirical analyses, respectively for dynamic characterization and fragility modeling. Using vibration time series records from 206 Reinforced Concrete (RC) buildings, period-height formulas are derived for different building subclasses. Uncertainties in period-height relationships are also quantified using Bayesian approach. Furthermore, dynamic characterization of neoclassical masonry structures is performed using recorded time series under damaged and retrofitted states, and evolution of dynamic characteristics under changed stiffness conditions are quantified to successfully demonstrate structural condition assessment using system identification. Deploying nonparametric and parametric system identification techniques, dynamic characteristics are estimated from ambient vibration time series records in all cases. In terms of the predictive seismic vulnerability formulation, seismic fragility functions for global and nonstructural component damage are derived using empirical database. Uncertainties in fragility models are quantified using Bayesian approach, implementing Markov Chain Monte Carlo (MCMC) simulation. Low-rise engineered and nonengineered RC are considered as two building classes to derive fragility functions, and global as well as component fragility models are derived for both building classes. The originality of the present study lies in uncertainty quantification in period-height relationships, characterization of dynamic characteristics in complex structures under changed states, prior effect quantification in fragility modeling, deployment of building-to-building variability through novel Beta-Binomial likelihood, and uncertainty quantification in seismic fragility functions for global and component damage.