Séminaire de Oindrila Kanjilal

Engineering Safe Structures : Advanced Monte Carlo Approaches

Predicting structural reliability increasingly requires quantifying inherent uncertainties while effectively integrating available data. While standard Monte Carlo simulation provides a robust benchmark for stochastic structural analysis, its computational cost becomes prohibitive when estimating the rare failure probabilities typical of critical infrastructures. This talk highlights advanced Monte Carlo methodologies engineered to overcome this computational bottleneck. We focus on adaptive importance-sampling formulations developed for efficient first-passage probability estimation under dynamic loads. We explore extensions of this framework to Bayesian model updating and reliability-based optimization, demonstrating how these specialized sampling schemes evaluate marginal likelihoods and update structural performance metrics as observational data are assimilated. Attendees will gain an understanding of how advanced sampling bridges physics-based computational models and operational monitoring data to support risk-informed structural decision-making.