ABSTRACT:  Dynamic effects dominate the operating conditions of many engineering structures. Therefore to improve the structural dynamic performance, in many areas of aeronautical and automotive engineering there has been a widespread use of virtual prototypes which can capture the underlying physics of the problem, and can be used to assess operating conditions not reproducible with experiments. Nonetheless, several parameters of a computational model such as geometry, material properties, loadings, boundary conditions, and structural joints required to investigate the behaviour of built-up structures may not be known precisely or may be affected by inherent variability arising from the manufacturing process. As a result, the performance of such engineering systems is uncertain, and it must include an estimation of the probability that the design targets will be met.

During this talk I will give an overview of approaches recently developed for Uncertainty Quantification in Structural Dynamics. In particular, I will talk about a non-parametric probabilistic approach that can be applied to linear high-frequency problems, the so-called Statistical Energy Analysis (SEA). I will then describe an approach combining the SEA non-parametric description parametric with parametric uncertainty models. Finally, I will describe a recently developed imprecise probability model based on Maximum Entropy concepts, and its application to random vibration problems. Recent developments on design decisions under imprecise probabilities are going to be briefly presented.