Computation of bridge seismic fragility by large-scale simulation for probabilistic resilience analysis

Seismic resilience of structures and infrastructure systems is a fast developing concept in the field of disaster management, promoting communities that are resistant and quickly recoverable in case of an extreme event. In this contest, probabilistic seismic demand and fragility analyses are two key elements of the seismic resilience assessment in the majority of the proposed methodologies. Several techniques are available to calculate fragility curves for different types of structures. In particular, to assess the seismic performance of the regional transportation infrastructure, methods for the fragility curve estimation for entire classes of bridges are required. These methods usually rely on a set of assumptions, partially because of the limited information. Other assumptions were introduced at the time when computational resources were inadequate for a purely numerical approach and closed-form solutions were a convenient alternative. For instance, some of these popular assumptions are aimed at simplifying the model of the engineering demand. In this paper, a simulation-based methodology is proposed, to take advantage of the computational resources widely available today and avoid such assumptions on the demand. The resulting increase in accuracy is estimated on a typical class of bridges (multi-span simply supported). Most importantly, the quantitative impact of the assumptions is assessed in the context of a life-cycle loss estimation analysis and resilience analysis. The results show that some assumptions preserve an acceptable level of accuracy, but others introduce a considerable error in the fragility curves and, in turn, in the expected resilience and life-cycle losses of the structure. Copyright (c) 2015 John Wiley & Sons, Ltd.