Statistical Analysis of Fragility Curves of Confined Masonry Walls

Masonry buildings are one of the most vulnerable structures against earthquakes, and therefore their seismic response should be evaluated to mitigate casualties in the future. To assess vulnerability and select retrofit approach, fragility curves can be utilized as powerful tools. Building fragility describes the likelihood of damage to a building due to random ground motions. Conventional methods for showing building fragilities are based on statistical analysis on the building models. Consequently, these analyses can be shown as fragility curves. These diagrams show the probability of exceeding a specific state of damage versus seismic intensity parameters. One of the most important parts of seismic assessment of building by using fragility curves is distribution function selection. It is common to use the two-parameter lognormal distribution in fragility analysis. However, evaluation of the other distribution functions can be useful to find a distribution function, which has the best fitting to damage indices. In this paper, the vulnerability of Confined Masonry (CM) walls, evaluated based on Park-Ang damage index by using the model developed in OPENSEES software in previous studies, has been used. The damage index values have been selected based on nonlinear dynamic analysis under thirty earthquake records to reduce the uncertainty of ground motion. Also, the goodness of fit of some distribution functions like normal, lognormal, Gumbel, etc. to the damage indices has been assessed. Finally fragility curves are obtained for two soil types B and C based selected distribution. The results show that the lognormal distribution has better fit in low levels of intensities. However, Gumbel distribution is better in higher levels. Fragility curves show the damage probability for various earthquakes intensities. In low intensity levels damage probability on soil type B is dominant. However, in higher intensities CM walls are damaged on soil type C severely.