Abstract

VAZQUEZ-GUILLEN, Felipe  and  AUVINET-GUICHARD, Gabriel. Component structural fragility facing correlated non-Gaussian uncertainties. Ing. invest. y tecnol. [online]. 2018, vol.19, n.1, pp.101-112. ISSN 2594-0732.  https://doi.org/10.22201/fi.25940732e.2018.19n1.009.

The modifications implemented on two classical methodologies of structural reliability that will allow to consider correlated non-Gaussian uncertainties at evaluating the structural fragility of components are described in this paper. The first methodology is based on the Monte Carlo method and the second is based on the First Order Reliability Method. Both methods are then used to construct fragility curves for the surge only overtopping erosion failure of one hypothetical levee. Results of fragility analysis show that: 1) Considering non-Gaussian marginal densities for the design variables, fragility curves differs significantly from the lognormal distribution; 2) The type of the random variable with the strongest influence on the levee failure does not necessarily determine the type of fragility curve. Therefore, within a more realistic context of non-Gaussian uncertainties, it is not correct to assume a prior theoretical model for the fragility curve ignoring the distributions of the basic variables because it is not possible to anticipate its type based only on the distribution of the random variable with the strongest incidence on the component failure.

Keywords : fragility curves; failure probability; overtopping erosion failure; Monte Carlo method; FORM method.

        · abstract in Spanish     · text in Spanish     · Spanish ( pdf )