System Identification of Constructed Facilities: Challenges and Opportunities Across Hazards Authors: Tracy Kijewski-Correa, University of Notre Dame, Notre Dame, IN, tkijewsk@nd.edu Ertugrul Taciroglu, University of California, Los Angeles, CA, etacir@ucla.edu James L. Beck, California Institute of Technology, Pasadena, CA, jimbeck@caltech.edu ABSTRACT The motivation, success and prevalence of full-scale monitoring of constructed buildings vary considerably across the hazard of concern (earthquakes, strong winds, etc.), due in part to various fiscal and life safety motivators. Yet while the challenges of successful deployment and operation of large-scale monitoring initiatives are significant, they are perhaps dwarfed by the challenges of data management, interrogation and ultimately system identification. Practical constraints on everything from sensor density to the availability of measured input has driven the development of a wide array of system identification and damage detection techniques, which in many cases become hazard-specific. In this study, the authors share their experiences in full- scale monitoring of buildings across hazards and the associated challenges of system identification. The study will conclude with a brief agenda for next generation research in the area of system identification of constructed facilities. INTRODUCTION The monitoring of constructed facilities has historically been tied to the need to understand in- situ behaviors, often spurred by suspect performance, as in the case of the infamous Boston Hancock Tower. The prevalence of full-scale monitoring in subsequent decades was largely tied, at least in Asia, to the proliferation of auxiliary damping devices, where sensors served a variety of purposes from actual feedback mechanisms in active control to a means to document the performance of these supplementary devices during earthquakes and typhoons. In parallel, monitoring efforts in seismically active zones within the United States received similar emphasis at this time for enhanced understanding of overall behavior and performance, while the wind engineering community turned its full-scale efforts toward developing databases of in-situ dynamic properties. Today, while monitoring efforts worldwide continue the task of in-situ validation of dynamic properties, dynamic load effects and response characteristics, added emphasis on rapid assessment and evaluation in the larger venue of structural health monitoring (SHM) has significantly expanded the literature on system identification (SI), though applications to actual structures are still quite limited. One obvious goal of most monitoring efforts is the identification of in-situ dynamic characteristics such as periods of vibration, mode shapes and critical damping ratios or in some cases, the direct stiffness and damping matrices themselves. A diverse cross section of approaches has been formulated to achieve these objectives, which can be broadly categorized by the hazard they address (earthquakes vs. strong winds). Thus it is worthwhile to underscore the