Abstract
Classical contingency analysis assesses the robustness of infrastructure systems by removing one component at the time (sometimes up to two components) and evaluating the effect on the system performance Y. In systems with dependent component failures, this approach might not identify critical system failure scenarios, e.g. if failures are caused by natural hazards or other common causes. In this contribution, we develop an approach to identify representative scenarios ST of component failures that are associated with damages or performance losses of a specific return period T. In practice, such scenarios are mostly defined based on historical data and expert knowledge, which often reflect past events but might not be representative of future events. Our approach is based on an initial Monte Carlo analysis of the system, resulting in an annual exceedance probability function of the system performance Y. Samples that approximately correspond to the value of Y associated with the return period of interest T are selected. The representative scenario for T is then identified by means of a clustering algorithm applied to these samples. The approach is demonstrated on a numerical example.
Original language | English |
---|---|
State | Published - 2019 |
Event | 13th International Conference on Applications of Statistics and Probability in Civil Engineering, ICASP 2019 - Seoul, Korea, Republic of Duration: 26 May 2019 → 30 May 2019 |
Conference
Conference | 13th International Conference on Applications of Statistics and Probability in Civil Engineering, ICASP 2019 |
---|---|
Country/Territory | Korea, Republic of |
City | Seoul |
Period | 26/05/19 → 30/05/19 |