Proceedings of 2009 NSF Engineering Research and Innovation Conference, Honolulu, Hawaii Grant # CMMI-0800399 Earthquake Damage Detection in Buildings and Early Warning Based on Wave Travel Times Maria I. Todorovska University of Southern California Abstract: This paper describes a novel method for structural health monitoring of buildings, currently being developed at the University of Southern California for future applications in earthquake early warning systems. The method uses data from vibrational sensors and is based on detecting changes in travel times of waves propagating through the structure. It is an intermediate scale method, which can point to the part of the structure that has been damaged using data from relatively small number of sensors. The method is illustrated using earthquake response data in two buildings damaged by earthquakes. The benefits to society from early warning systems of damage in structures, before physical inspection is possible, are discussed, both from the point of view of safety and from the point of view of avoiding needless service interruption. 1. Introduction: Structural health monitoring aims to determine and track the structural integrity and to assess the nature of damage. The ability to monitor remotely the health of an instrumented structure, detect damage as it occurs, and issue an early warning after an extreme event (earthquake, explosion or some other natural or man made disaster), before physical inspection is possible, has immense potential benefits in reducing loss of life and injuries, in emergency response, and in recovery following the disaster. For example, a building owner can make a timely decision on evacuation of an unsafe building, reducing the risk of loss of life and injuries caused by potential collapse of the weakened structure during shaking from aftershocks. Similarly, a bridge operator can make a timely decision to close the bridge, and redirect the traffic. Another important benefit from structural health monitoring is detecting hidden damage (its precise location and extent), due to an extreme event or gradual decay in time, without a costly physical inspection of every structural member. The development of such methodology that would work in practical real life applications, however, is a very difficult task, and, considering the challenges faced and the potential benefits, has the elements of a grand challenge problem in civil engineering (Farrar and Worden 2007). To be practically useful, the structural health monitoring systems need to be robust, reliable and sufficiently sensitive when applied to actual earthquake data. They should neither miss to detect significant structural damage, nor create false alarms leading to needless and costly service interruptions. This is particularly important for critical facilities, such as hospitals, and nuclear power plants. In addition, they should be able to point out to the location of the damage. Despite the progress made in this field in more than 30 years of research, currently there are no such systems. This paper describes a promising new structural health monitoring method currently being developed at the Strong Motion Laboratory of the University of Southern California. The method uses data from vibrational sensors is based on detecting changes in structural stiffness based on analysis of travel times of seismic waves propagating through the structure. The method is described within the context of the vibrational methods, which are briefly reviewed. The advantages of the new wave method over the prevailing modal methods are discussed. The wave methods of analysis of building response are also reviewed. The structural health monitoring methods is illustrated by results from proof of concept applications to two buildings damaged by earthquakes – the Imperial County Services Building and the Van Nuys 7-story hotel (Todorovska and Trifunac 2008a,b). This method is being further developed and calibrated using earthquake response data recorded in full-scale building. A review of recent developments in structural health monitoring, as applied to civil and mechanical systems, can be found in Chang et al. (2003) and Liu et al. (2006). The earliest, and still by far the most widely used methods for civil structures are those that use data from vibrational sensors (accelerometers or velocity meters), and detect changes in the vibrational characteristics of the structure (frequencies of vibration and mode shapes, Doebling et al. 1996, Sohn et al. 2003, Carden and Fanning 2004, Farrar and Worden 2007, Todorovska 2009c). NSF GRANT # 0800399 NSF PROGRAM NAME: Civil, Mechanical and Manufacturing Innovation