MS 28

Imprecision and Randomness in Dynamical Systems

Organizers

Matthias Faes,
TU Dortmund University, Germany

Alice Cicirello,
University of Cambridge, UK

Alba Sofi,
University Mediterranea of Reggio Calabria, Italy

Sifeng Bi,
University of Southampton, UK

Marcos Valdebenito,
TU Dortmund University, Germany

Chao Jiang,
Hunan University, China

Abstract

Systems and structures subject to dynamic loading are usually affected by different sources of uncertainty. For example, the presence of uncertainties in time-varying operational and environmental loading as well as the system’s time-varying properties and boundary conditions (for example due to degradation phenomena) can severely affect the uncertainty in the system’s response. Classical probabilistic tools offer an excellent means for characterizing dynamic behaviour whenever randomness is present, especially when the system parameters and inputs are assumed time-invariant. However, for those cases where the source of uncertainty is imprecision due to issues such as lack of knowledge or scarce data, non-traditional methods for uncertainty quantification may offer an excellent complement to probabilistic methods. Yet, the application of probabilistic and non-traditional methods for uncertainty quantification for dynamical problems becomes a daunting task, as it is necessary to keep track of both sources of uncertainty without mixing them, often while dealing with an expensive-to-evaluate physics-based model, and time-varying uncertainty descriptions. Therefore, the aim of this MS is to bring together some of the latest developments on methods for coping with imprecision and randomness in dynamics, including (but not limited to):

  • Different models for representing uncertainty such as, e.g. p-boxes, imprecise stochastic processes, distribution-free models, etc. and their application for describing time-varying loads and time-varying system’s properties.

  • Novel formulations for coping with aleatoric and epistemic uncertainty, such as advanced simulation methods, surrogate models, etc.

  • Inverse uncertainty quantification in the presence of aleatoric and epistemic uncertainty and sparse data and/or limited knowledge, with emphasis on stochastic model updating, model verification and validation.

  • Practical applications involving challenging problems in system’s dynamics uncertainty quantification.