Importance Sampling in Life-Cycle Seismic Fragility and Risk Assessment of Aging Bridge Networks

Life-cycle structural reliability of deteriorating systems and multi-hazard risk assessment of aging infrastructure networks involve complex time-variant processes characterized by impactful uncertainties. Simulation methods are frequently the only viable tools to accurately estimate time-variant failure probabilities and risk metrics. However, simulation-based techniques are time-consuming and might be computationally inefficient and unfeasible in practice, particularly when analysis of large-scale systems are required to assess numerically sensitive performance indicators. This paper proposes a novel computational approach based on Importance Sampling to efficiently estimate the time-variant seismic risk of aging road networks. In the proposed methodology, the seismic capacity of deteriorating structural systems is efficiently simulated to account for the time-variant model uncertainties typical of life-cycle structural reliability problems. The possible improved trade-off in terms of sample size and estimate accuracy is tested in comparison with traditional Monte Carlo simulation approaches based on a practical application concerning the life-cycle seismic risk assessment of a road network with spatially-distributed deteriorating vulnerable bridges