ABSTRACT
This paper presents a mathematical framework to investigate the relative importance of seismic damage scenarios to the probability of failing to meet target performance measures for spatially-distributed aging bridge networks. The proposed framework relies on a statistical approach adapted from disaggregation procedures typical of probabilistic seismic hazard analysis. The estimate accuracy is enhanced using a novel simulation-based methodology proposed in previous works based on Importance Sampling with Stationary Proposal distributions (SP-IS). The basic random variables involved in risk assessment are efficiently sampled from a near-optimal simulation density based on the minimization of the Kullback–Leibler cross-entropy. A simple road network is investigated to highlight the benefits to computational effort of the proposed SP-IS numerical method and to explore the potentialities of the concept of damage disaggregation in communicating to infrastructure managers and policy makers the large-scale consequences of natural hazards and aid the optimal management and prioritization of essential maintenance interventions.
