Risk and hazard assessment of extreme natural events for critical infrastructures

Abstract

Human society is frequently surprised by extreme natural events that lead to tremendous losses both in human lives as well as in economic capital. Risk assessment methods regularly fail to predict such disasters by treating extreme events as low probability, tolerable ones. Critical infrastructures and life lines are often found to be poorly protected due to their inadequate design basis. A method of hazard assessment for extreme natural events is presented that allows for the consideration of unexpected extreme events (‘black swan’ events). Including such events into the design basis prevents the failure of critical infrastructures due to ‘cliff-edge’ effects. The method makes use of some general properties of heavy-tailed distributions and the mathematical theory of records, and takes advantage of a distribution-free approach without need to calculate probabilities of exceedance. The application of the method is demonstrated on several examples for nuclear power plants (NPP), including High Wind and High Temperature hazards. The results are compared with the results of hazard assessment using conventional probabilistic hazard analysis methods. The risk implications are discussed.