Organizers
Every numerical and computational prediction is characterized by several sources of uncertainty arising from model assumptions and simplifications of the reality, and from the uncertainty about input parameters (i.e. physical and mechanical properties, geometries, boundary conditions). Each model can be described by a forward problem, which predicts some quantities of interest of the system given a set of unknown/uncertain parameters, and by the corresponding inverse problem, which consists in estimating the set of these parameters from a set of measured/observed data. In realistic applications, the data are often noisy and/or incomplete and this introduces a further level of uncertainty. Bayesian inference represents an actual effective tool to quantify the uncertainty around models’ input parameters, but novel mathematical formulations and sampling techniques need to be developed. The objective of this mini-symposium is to highlight the new research trends in the field of uncertainty quantification methods in structural engineering applications with a special focus on surrogate models, reduced order models and Monte Carlo methods. Moreover, contributions presenting novel applications of Bayesian inference and/or methods for the inverse problem solutions (i.e. structural identification, surrogate model based sampling techniques, structural health monitoring applications) are welcome.
- Chiara Pepi, PhD, Department of Civil and Environmental Engineering, University of Perugia, Perugia, Italy, chiara.pepi@unipg.it
- Silvia Monchetti, PhD, Department of Civil and Environmental Engineering, University of Florence, Florence, Italy, silvia.monchetti@unifi.it
- Viscardi Cecilia, PhD, Department of Statistics, Computer Science, Applications, University of Florence, Florence, Italy, cecilia.viscardi@unifi.it
Every numerical and computational prediction is characterized by several sources of uncertainty arising from model assumptions and simplifications of the reality, and from the uncertainty about input parameters (i.e. physical and mechanical properties, geometries, boundary conditions). Each model can be described by a forward problem, which predicts some quantities of interest of the system given a set of unknown/uncertain parameters, and by the corresponding inverse problem, which consists in estimating the set of these parameters from a set of measured/observed data. In realistic applications, the data are often noisy and/or incomplete and this introduces a further level of uncertainty. Bayesian inference represents an actual effective tool to quantify the uncertainty around models’ input parameters, but novel mathematical formulations and sampling techniques need to be developed. The objective of this mini-symposium is to highlight the new research trends in the field of uncertainty quantification methods in structural engineering applications with a special focus on surrogate models, reduced order models and Monte Carlo methods. Moreover, contributions presenting novel applications of Bayesian inference and/or methods for the inverse problem solutions (i.e. structural identification, surrogate model based sampling techniques, structural health monitoring applications) are welcome.
