Estimation uncertainty for some common seismic fragility curve fitting methods

This technical note illustrates and makes available some simple procedures to assess the estimation uncertainty for the parameters of seismic fragility curves. The considered fragility fitting methods refer to the lognormal assumption and are supposed to be based on the results of multi-stripe dynamic analysis of a deterministic nonlinear structural model, so that the uncertainty in the fragility parameters arises from the so-called record-to record variability. The discussed procedures are based on the statistics approach of resampling with substitution, which is commonly referred to as bootstrap. It is also briefly discussed how the estimation uncertainty depends on the maximum value of the probability of failure given seismic intensity that is observed from structural analysis. This work may aid earthquake engineering practice because, both the curve fitting and estimation uncertainty algorithms are implemented in a major update of an application-ready software tool made available at https://www.reluis.it/it/progettazione/software/r2r-eu.html.

Automated summary

This technical note illustrates simple procedures for assessing estimation uncertainty of seismic fragility curve parameters based on the lognormal assumption and multi-stripe dynamic analysis results. The discussed procedures are based on statistics resampling with substitution, commonly referred to as bootstrap, and how the estimation uncertainty depends on the maximum value of the probability of failure given seismic intensity observed from structural analysis.