Introducing Correlation among Fragility Functions for Multiple Components

Abstract eng:
Correlations are known to be an important consideration when obtaining accurate probabilistic seismic loss estimation results. Correlations among structural response values are straightforward to measure and incorporate, but it is difficult to incorporate correlations in the element damage predicted by fragility functions. In the most widely used component-based loss estimation procedures (e.g., [1-5]), structural response correlations are incorporated, but typically only the special cases of independent or perfectly-correlated component damage states are allowed. A simple method for considering partially correlated damage is proposed here. Whereas traditional calculations first compute a probability of damage and then consider occurrence of that damage to be random, the method proposed here instead considers the “capacity” of the components to be random, which allows partial correlation of capacities to be easily introduced (because the random component capacity is continuous, unlike the discrete damage states). The approach relies on Monte Carlo simulation, making it compatible with the Monte Carlo analysis approaches used within some existing loss assessment procedures, although purely analytic implementations are also feasible.

• Export as: BibTeX | MARC | MARCXML | DC | EndNote | NLM | RefWorks
• Add to your basket