Use of Kriging metamodels for seismic fragility analysis of structures - Archive ouverte HAL Accéder directement au contenu
Chapitre D'ouvrage Année : 2020

Use of Kriging metamodels for seismic fragility analysis of structures

Résumé

In civil engineering, a seismic fragility curve is popularly used to predict failure probability of structures under different earthquakes, and hence propose essential rehabilitation strategies through risk assessment for future earthquakes. The curve shows the failure probability as a function of seismic intensity, e.g., spectral acceleration at fundamental frequencies of structures (Sa, T1), and can be obtained using one of three approaches: engineering judgment, empirical studies or numerical simulations. The paper focuses on constructing seismic fragility curves using numerical simulations, where robust approaches of seismic reliability analysis are based on direct Monte Carlo simulation technique. The MCS based method usually requires a relatively large number of simulations to obtain a sufficiently reliable estimate of the fragility. It therefore becomes computationally expensive and time consuming as generating the simulations using the actual model or called full model of the structure. In this regard, this paper suggests using Kriging metamodel as a viable alternative of the actual model to reduce computational costs in seismic fragility computation. The Kriging metamodel is constructed based on the training samples of input and corresponding output responses of the structure. The validation of this method is performed on two numerical examples.
Fichier non déposé

Dates et versions

hal-02334886 , version 1 (27-10-2019)

Identifiants

Citer

Cong-Thuat Dang, Thanh Hai Tran, Duy-My Nguyen, My Pham, Thien-Phu Le. Use of Kriging metamodels for seismic fragility analysis of structures. Lecture Notes in Civil Engineering, 54, pp.299--304, 2020, ⟨10.1007/978-981-15-0802-8_45⟩. ⟨hal-02334886⟩
160 Consultations
0 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More