Recent earthquakes worldwide underscore the urgent need for accurate seismic response prediction of concrete gravity dams, where failure is unacceptable. However, assessing dam response to strong ground motions remains challenging due to uncertainties in seismic input, material properties, and the limited availability of experimental data for model validation.This study presents a 2D linear finite element (FE) model of the Pine Flat Dam, a concrete gravity dam in California. The model incorporates key physical mechanisms, including water compressibility, wave absorption, dam-foundation interaction, and radiation damping. Bayesian model updating is performed using Transitional Markov Chain Monte Carlo under two distinct load scenarios: (1) Variations in reservoir water elevation, where quasi-static displacements inform parameter estimation, and (2) Earthquake ground motion excitation, where acceleration and hydrodynamic pressure measurements provide heterogenous data for parameter inference. By leveraging these distinct measurement types, the study examines parameter estimation consistency and the extent to which each dataset informs the inference of material properties, contributing to a more robust assessment of model parameter uncertainty.
Bayesian Model Updating of a Dam-Water-Foundation Rock System: A Case Study of Pine Flat Dam
Camata G.;
2025-01-01
Abstract
Recent earthquakes worldwide underscore the urgent need for accurate seismic response prediction of concrete gravity dams, where failure is unacceptable. However, assessing dam response to strong ground motions remains challenging due to uncertainties in seismic input, material properties, and the limited availability of experimental data for model validation.This study presents a 2D linear finite element (FE) model of the Pine Flat Dam, a concrete gravity dam in California. The model incorporates key physical mechanisms, including water compressibility, wave absorption, dam-foundation interaction, and radiation damping. Bayesian model updating is performed using Transitional Markov Chain Monte Carlo under two distinct load scenarios: (1) Variations in reservoir water elevation, where quasi-static displacements inform parameter estimation, and (2) Earthquake ground motion excitation, where acceleration and hydrodynamic pressure measurements provide heterogenous data for parameter inference. By leveraging these distinct measurement types, the study examines parameter estimation consistency and the extent to which each dataset informs the inference of material properties, contributing to a more robust assessment of model parameter uncertainty.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
