Fragility analysis for the bridge control benchmark problem

This study evaluates the fragility curves of a cable stayed bridge using Monte Carlo simulations. The presentation of the vulnerability information in the form of fragility curves is a widely practiced approach when several uncertain sources are involved. Seismic fragility is the probability of system failure as a function of some seismic intensity parameters and failure is defined as the exceedance of some limit states by the corresponding structural response quantities. The limit states for fragility analysis are considered as random variables and a multidimensional threshold of limit states has been used [5] to calculate fragility information of structural systems that consider multiple variables. The benchmark problem of a cable stayed bridge is considered as a case study. The bridge is located in the USA spamming the Mississippi river near Cape Girardeau, and is a part of the benchmark problem phase II that, with respect to phase I, includes the bi-directional effects of the seismic action. The bridge is located in the New Madrid seismic zone (Cape Girardeau: Lat. 37.2971N Long. -89.5163W) for which no strong motion records from similar historical earthquakes exists. Hence synthetic accelerograms are used. The synthetic accelerograms are generated by the program SMSIM_TD version 2.10 [2]. A passive control technique is adopted and passive control devices were modelled using the Bouc-Wen endochronic hysteretic model. Results are compared in term of fragility curves. The reference variables in building the fragility curves are the moments and the shears at the base of the piers and at the deck level respectively. The tension in the cable has also been used as well as the displacements of the deck. The fragility curves evaluated for the passive control techniques are compared with the fragility curves estimated for the uncontrolled case. In order to consider the uncertainties related to the ground motion, four different hazard levels are considered, while the uncertainties in the structural characteristics are introduced by defining the different performance thresholds as random variables assuming a log-normal distribution. The fragility evaluation shows how important a correct estimation of the limit state is for the comparison of different retrofit techniques.