.png){kind=icon}
.png){kind=icon}
Return
Seismic fragility analysis of underground structures using Bayesian updated bilinear seismic demand models
.png){kind=icon}
PRE
AI
W
M
P
DOI：10.1016/j.ress.2025.112125.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Seismic fragility analysis is critical to performance-based seismic design for structures. This study proposes a Bayesian inference-based method for seismic fragility analysis of underground structures, incorporating the spatial variability of soil parameters using the Karhunen-Loève expansion. The proposed method employs Bayesian inference to update a bilinear seismic demand model and uses the Metropolis–Hastings algorithm to obtain posterior samples of the fitted parameters, comprehensively accounting for the uncertainties arising from ground motions, soil parameters, and the seismic demand model itself. The fragility curves of a subway station structure with the spatially variable soil parameters were constructed. The results indicate that neglecting soil spatial variability leads to an underestimation of structural damage probabilities under low seismic intensities and an overestimation under high seismic intensities, with this effect being more pronounced when the bilinear model is adopted. Furthermore, the fragility curves based on linear and bilinear seismic demand models were compared. It was found that the seismic intensity and demand data for this case exhibit significant nonlinearity in logarithmic space, which cannot be effectively captured using a linear demand model. A linear seismic demand model overestimates the probabilities of minor damage and moderate damage and underestimates the probabilities of extensive damage and collapse. The proposed method provides a precise analytical framework for seismic fragility evaluation of underground structures.
Journal
R
IF:
11
Papers:
9.0K
Citations:
4.2W
Organization
Cited Papers