Post‐Hazard learning, risk assessment and decision‐making for infrastructure systems
Armen Der Kiureghian
Department of Civil & Environmental Engineering
University of California, Berkeley
Robust performance and rapid recovery of infrastructure systems in the immediate aftermath of a major hazard are crucial for mitigating losses and assuring well‐being of communities. Infrastructures, such as transportation and communication networks and power, water and gas distribution systems, are especially vulnerable to natural and man‐made hazards due to their spatially distributed exposure, interdependence between components, and multiplicity of failure modes. In this lecture, I will use the Bayesian network methodology to model the hazard and the infrastructure system and to process information gained from sensors, and I will use influence diagrams to make decisions on operational levels of system components and to prioritize component inspections. An application to a hypothetical model of the California high‐speed rail system in the aftermath of an earthquake in the Bay Area will demonstrate the main ideas of the approach.